Bipolar Disorder

Ms. K, age 35, soon will deliver her sec­ond child. She has a 12-year history of bipolar disorder, which was well controlled with lithium, 1,200 mg/d. During her first pregnancy 3 years ago, Ms. K stopped tak­ing lithium because she was concerned about the risk of Ebstein’s anomaly. She experienced a bipolar relapse after her healthy baby was born, and developed postpartum psychosis that was treated by restarting lithium, 1,200 mg/d, and adding olanzapine, 10 mg/d.

Ms. K has continued these medications throughout her current pregnancy. She wants to breast-feed her infant and is concerned about the effects that psychotropics might have on her newborn.

Breast-feeding and medications

The benefits of breast-feeding for mother and infant are well-known. Despite this, some women with bipolar disorder are advised not to breast-feed or, worse, to discontinue their medications in order to breast-feed. Decisions about breast-feeding while taking medications should be based on evidence of benefits and risks to the infant, along with a discussion of the risks of untreated illness, which is high postpartum. The prescribing information for many of the medications used to treat bipolar disorder advise against breast-feeding, although there is little evidence of harm.

Drug dosages and levels in breast milk can be reported a few different ways:
   • percentage of maternal dosage mea­sured in the breast milk
   • percentage weight-adjusted maternal dosage
   • percentage of maternal plasma level, and milk-to-plasma ratio (M:P).

Daily infant dosage can be calculated by multiplying the average concentration of the drug in breast milk (mg/mL) by the average volume of milk the baby ingests in 24 hours (usually 150 mL).¹ The rela­tive infant dosage can be calculated as the percentage maternal dosage, which is the daily infant dosage (mg/kg/d) ÷ maternal dosage (mg/kg/d) × 100.¹

According to the American Academy of Pediatrics, ≤10% of the maternal dos­age is compatible with breast-feeding.¹ Most psychotropics studied fall below this threshold. Keep in mind that all published research is for breast-feeding a full-term infant; exercise caution with premature or low birth weight infants. Infants born to mothers taking a psychotropic should be monitored for withdrawal symptoms, which might be associated with antide­pressants and benzodiazepines, but other­wise are rare.

Lithium
Breast-feeding during lithium treatment has been considered contraindicated based on early reports that lithium was highly excreted in breast milk.² A 2003 study² of 11 women found that lithium was excreted in breast milk in amounts between zero and 30% of maternal dosage (mean, 12.2% ± 8.5%; median, 11.2%; 95% CI, 6.3% to 18.0%). Researchers measured serum con­centrations in 2 infants and found that 1 received 17% to 20% of the maternal dos­age, and the other showed 50%. None of the infants experienced adverse events. In a study of 10 mother-infant pairs, breast milk lithium concentration averaged 0.35 mEq/L (standard deviation [SD] = 0.10, range 0.19 to 0.48 mEq/L), with paired infant serum concentrations of 0.16 mEq/L (SD = 0.06, range 0.09 to 0.25 mEq/L).³ Some transient abnormali­ties were found in infant serum concen­trations of thyroid-stimulating hormone (TSH), blood urea nitrogen, and creatinine; there were no adverse effects on develop­ment. The authors recommend monitoring for TSH abnormalities in infants.

Olanzapine
Olanzapine prescribing information cites a study reporting that 1.8% of the maternal dosage is transferred to breast milk.⁴ Yet the olanzapine prescribing information states, “It is recommended that women receiv­ing olanzapine should not breast-feed.” Olanzapine use during breast-feeding has been studied more than many medications, in part because of a database maintained by the manufacturer. In a study using the man­ufacturer’s database (N = 102) adverse reac­tions were reported in 15.6% of the infants, with the most common being somnolence (3.9%), irritability (2%), tremor (2%), and insomnia (2%).⁵

Other second-generation antipsychotics
Aripiprazole. The only case report of aripip­razole excretion in human breast milk found a concentration of approximately 20% of the maternal plasma level and an M:P ratio of 0.18:0.2.⁶

Asenapine. According to asenapine pre­scribing information⁷ and a literature search, it is not known whether asenapine is excreted in breast milk of humans, although it is found in the milk of lactating rats.

Lurasidone. According to the lurasidone prescribing information⁸ and a literature search, it is not known whether lurasidone is excreted in human breast milk, although it is found in the milk of lactating rats.

Quetiapine. An initial study reported that 0.09% to 0.43% of the maternal dosage of quetiapine was excreted in breast milk.⁹ Further studies found excretion to be 0.09% of maternal dosage, with infant plasma levels reaching 6% of the maternal dos­age.¹⁰ A case series found that one-third of babies exposed to quetiapine during breast-feeding showed some neurodevelopmental delay, although these mothers also were taking other psychotropics.¹¹

Risperidone. A 2000 study¹² of risperidone in lactation reported that 0.84% weight-adjusted maternal risperidone dosage and 3.46% of its metabolite 9-hydroxyrisperi­done is transferred to the infant. A later study showed 2.3% to 4.7% of the mater­nal dosage is transferred, with no adverse events reported in infants.¹³ A case study reported no adverse events and normal neu­rodevelopment in a the child of a mother taking risperidone.¹⁴

Ziprasidone. According to the ziprasidone prescribing information¹⁵ and a literature search, is not known whether ziprasidone is excreted in human breast milk.

See the Table^{4,6-10,12,13,15} for a summary of the evidence levels of second-generation antipsychotics that are excreted in breast milk.

Valproic acid. Because of its high plasma protein binding, valproic acid does not pass readily into the breast milk. Newborns receive approximately 1.4% to 1.7% of the maternal dosage.¹⁶ Caution is advised, however, because of some reported adverse events. One case reported thrombocytopenic purpura and anemia in an infant.¹⁹ Valproic acid is considered to be compatible with breast-feeding with proper monitoring.

Benzodiazepines
Benzodiazepines can be helpful adjunctive medications to aid sleep, which is essen­tial for the mother’s and infant’s health. In a prospectively recruited, retrospec­tively assessed cohort study that evaluated 124 women taking benzodiazepines while breast-feeding, adverse effects, specifically sedation, were noted in 1.6% of infants.²⁰

Future developments in prescribing information
Under a 2008 FDA recommendation, the “Nursing Mothers” section of prescribing information would be replaced with a sec­tion entitled “Lactation.” This new head­ing would include the sub-headings Risk Summary, Clinical Considerations, and Data.¹ It is expected that this new format will be more practical and will help clini­cians and patients make informed deci­sions. The prescribing changes will be in effect on June 30, 2015.²¹

Related Resources
• Massachusetts General Hospital Center for Women’s Mental Health. www.womensmentalhealth.org.
• LactMed. http://toxnet.nlm.nih.gov/newtoxnet/lactmed.htm.
• MOTHERISK. www.motherisk.org/women/ breastfeeding.jsp.

Drug Brand Names
Aripiprazole • Abilify                    Olanzapine • Zyprexa
Asenapine • Saphris                    Quetiapine • Seroquel
Carbamazepine • Tegretol           Risperidone • Risperdal
Lamotrigine • Lamictal                 Valproic acid • Depakene
Lithium • Eskalith, Lithobid           Ziprasidone • Geodon
Lurasidone • Latuda

Disclosure
The author reports no financial relationships with any company whose products are mentioned in this article or with manufacturers of competing products.

Ms. K, age 35, soon will deliver her sec­ond child. She has a 12-year history of bipolar disorder, which was well controlled with lithium, 1,200 mg/d. During her first pregnancy 3 years ago, Ms. K stopped tak­ing lithium because she was concerned about the risk of Ebstein’s anomaly. She experienced a bipolar relapse after her healthy baby was born, and developed postpartum psychosis that was treated by restarting lithium, 1,200 mg/d, and adding olanzapine, 10 mg/d.

Ms. K has continued these medications throughout her current pregnancy. She wants to breast-feed her infant and is concerned about the effects that psychotropics might have on her newborn.

Breast-feeding and medications

The benefits of breast-feeding for mother and infant are well-known. Despite this, some women with bipolar disorder are advised not to breast-feed or, worse, to discontinue their medications in order to breast-feed. Decisions about breast-feeding while taking medications should be based on evidence of benefits and risks to the infant, along with a discussion of the risks of untreated illness, which is high postpartum. The prescribing information for many of the medications used to treat bipolar disorder advise against breast-feeding, although there is little evidence of harm.

Drug dosages and levels in breast milk can be reported a few different ways:
   • percentage of maternal dosage mea­sured in the breast milk
   • percentage weight-adjusted maternal dosage
   • percentage of maternal plasma level, and milk-to-plasma ratio (M:P).

Daily infant dosage can be calculated by multiplying the average concentration of the drug in breast milk (mg/mL) by the average volume of milk the baby ingests in 24 hours (usually 150 mL).¹ The rela­tive infant dosage can be calculated as the percentage maternal dosage, which is the daily infant dosage (mg/kg/d) ÷ maternal dosage (mg/kg/d) × 100.¹

According to the American Academy of Pediatrics, ≤10% of the maternal dos­age is compatible with breast-feeding.¹ Most psychotropics studied fall below this threshold. Keep in mind that all published research is for breast-feeding a full-term infant; exercise caution with premature or low birth weight infants. Infants born to mothers taking a psychotropic should be monitored for withdrawal symptoms, which might be associated with antide­pressants and benzodiazepines, but other­wise are rare.

Lithium
Breast-feeding during lithium treatment has been considered contraindicated based on early reports that lithium was highly excreted in breast milk.² A 2003 study² of 11 women found that lithium was excreted in breast milk in amounts between zero and 30% of maternal dosage (mean, 12.2% ± 8.5%; median, 11.2%; 95% CI, 6.3% to 18.0%). Researchers measured serum con­centrations in 2 infants and found that 1 received 17% to 20% of the maternal dos­age, and the other showed 50%. None of the infants experienced adverse events. In a study of 10 mother-infant pairs, breast milk lithium concentration averaged 0.35 mEq/L (standard deviation [SD] = 0.10, range 0.19 to 0.48 mEq/L), with paired infant serum concentrations of 0.16 mEq/L (SD = 0.06, range 0.09 to 0.25 mEq/L).³ Some transient abnormali­ties were found in infant serum concen­trations of thyroid-stimulating hormone (TSH), blood urea nitrogen, and creatinine; there were no adverse effects on develop­ment. The authors recommend monitoring for TSH abnormalities in infants.

Olanzapine
Olanzapine prescribing information cites a study reporting that 1.8% of the maternal dosage is transferred to breast milk.⁴ Yet the olanzapine prescribing information states, “It is recommended that women receiv­ing olanzapine should not breast-feed.” Olanzapine use during breast-feeding has been studied more than many medications, in part because of a database maintained by the manufacturer. In a study using the man­ufacturer’s database (N = 102) adverse reac­tions were reported in 15.6% of the infants, with the most common being somnolence (3.9%), irritability (2%), tremor (2%), and insomnia (2%).⁵

Other second-generation antipsychotics
Aripiprazole. The only case report of aripip­razole excretion in human breast milk found a concentration of approximately 20% of the maternal plasma level and an M:P ratio of 0.18:0.2.⁶

Asenapine. According to asenapine pre­scribing information⁷ and a literature search, it is not known whether asenapine is excreted in breast milk of humans, although it is found in the milk of lactating rats.

Lurasidone. According to the lurasidone prescribing information⁸ and a literature search, it is not known whether lurasidone is excreted in human breast milk, although it is found in the milk of lactating rats.

Quetiapine. An initial study reported that 0.09% to 0.43% of the maternal dosage of quetiapine was excreted in breast milk.⁹ Further studies found excretion to be 0.09% of maternal dosage, with infant plasma levels reaching 6% of the maternal dos­age.¹⁰ A case series found that one-third of babies exposed to quetiapine during breast-feeding showed some neurodevelopmental delay, although these mothers also were taking other psychotropics.¹¹

Risperidone. A 2000 study¹² of risperidone in lactation reported that 0.84% weight-adjusted maternal risperidone dosage and 3.46% of its metabolite 9-hydroxyrisperi­done is transferred to the infant. A later study showed 2.3% to 4.7% of the mater­nal dosage is transferred, with no adverse events reported in infants.¹³ A case study reported no adverse events and normal neu­rodevelopment in a the child of a mother taking risperidone.¹⁴

Ziprasidone. According to the ziprasidone prescribing information¹⁵ and a literature search, is not known whether ziprasidone is excreted in human breast milk.

See the Table^{4,6-10,12,13,15} for a summary of the evidence levels of second-generation antipsychotics that are excreted in breast milk.

Valproic acid. Because of its high plasma protein binding, valproic acid does not pass readily into the breast milk. Newborns receive approximately 1.4% to 1.7% of the maternal dosage.¹⁶ Caution is advised, however, because of some reported adverse events. One case reported thrombocytopenic purpura and anemia in an infant.¹⁹ Valproic acid is considered to be compatible with breast-feeding with proper monitoring.

Benzodiazepines
Benzodiazepines can be helpful adjunctive medications to aid sleep, which is essen­tial for the mother’s and infant’s health. In a prospectively recruited, retrospec­tively assessed cohort study that evaluated 124 women taking benzodiazepines while breast-feeding, adverse effects, specifically sedation, were noted in 1.6% of infants.²⁰

Future developments in prescribing information
Under a 2008 FDA recommendation, the “Nursing Mothers” section of prescribing information would be replaced with a sec­tion entitled “Lactation.” This new head­ing would include the sub-headings Risk Summary, Clinical Considerations, and Data.¹ It is expected that this new format will be more practical and will help clini­cians and patients make informed deci­sions. The prescribing changes will be in effect on June 30, 2015.²¹

Related Resources
• Massachusetts General Hospital Center for Women’s Mental Health. www.womensmentalhealth.org.
• LactMed. http://toxnet.nlm.nih.gov/newtoxnet/lactmed.htm.
• MOTHERISK. www.motherisk.org/women/ breastfeeding.jsp.

Drug Brand Names
Aripiprazole • Abilify                    Olanzapine • Zyprexa
Asenapine • Saphris                    Quetiapine • Seroquel
Carbamazepine • Tegretol           Risperidone • Risperdal
Lamotrigine • Lamictal                 Valproic acid • Depakene
Lithium • Eskalith, Lithobid           Ziprasidone • Geodon
Lurasidone • Latuda

Disclosure
The author reports no financial relationships with any company whose products are mentioned in this article or with manufacturers of competing products.

Ms. K, age 35, soon will deliver her sec­ond child. She has a 12-year history of bipolar disorder, which was well controlled with lithium, 1,200 mg/d. During her first pregnancy 3 years ago, Ms. K stopped tak­ing lithium because she was concerned about the risk of Ebstein’s anomaly. She experienced a bipolar relapse after her healthy baby was born, and developed postpartum psychosis that was treated by restarting lithium, 1,200 mg/d, and adding olanzapine, 10 mg/d.

Ms. K has continued these medications throughout her current pregnancy. She wants to breast-feed her infant and is concerned about the effects that psychotropics might have on her newborn.

Breast-feeding and medications

The benefits of breast-feeding for mother and infant are well-known. Despite this, some women with bipolar disorder are advised not to breast-feed or, worse, to discontinue their medications in order to breast-feed. Decisions about breast-feeding while taking medications should be based on evidence of benefits and risks to the infant, along with a discussion of the risks of untreated illness, which is high postpartum. The prescribing information for many of the medications used to treat bipolar disorder advise against breast-feeding, although there is little evidence of harm.

Drug dosages and levels in breast milk can be reported a few different ways:
   • percentage of maternal dosage mea­sured in the breast milk
   • percentage weight-adjusted maternal dosage
   • percentage of maternal plasma level, and milk-to-plasma ratio (M:P).

Daily infant dosage can be calculated by multiplying the average concentration of the drug in breast milk (mg/mL) by the average volume of milk the baby ingests in 24 hours (usually 150 mL).¹ The rela­tive infant dosage can be calculated as the percentage maternal dosage, which is the daily infant dosage (mg/kg/d) ÷ maternal dosage (mg/kg/d) × 100.¹

According to the American Academy of Pediatrics, ≤10% of the maternal dos­age is compatible with breast-feeding.¹ Most psychotropics studied fall below this threshold. Keep in mind that all published research is for breast-feeding a full-term infant; exercise caution with premature or low birth weight infants. Infants born to mothers taking a psychotropic should be monitored for withdrawal symptoms, which might be associated with antide­pressants and benzodiazepines, but other­wise are rare.

Lithium
Breast-feeding during lithium treatment has been considered contraindicated based on early reports that lithium was highly excreted in breast milk.² A 2003 study² of 11 women found that lithium was excreted in breast milk in amounts between zero and 30% of maternal dosage (mean, 12.2% ± 8.5%; median, 11.2%; 95% CI, 6.3% to 18.0%). Researchers measured serum con­centrations in 2 infants and found that 1 received 17% to 20% of the maternal dos­age, and the other showed 50%. None of the infants experienced adverse events. In a study of 10 mother-infant pairs, breast milk lithium concentration averaged 0.35 mEq/L (standard deviation [SD] = 0.10, range 0.19 to 0.48 mEq/L), with paired infant serum concentrations of 0.16 mEq/L (SD = 0.06, range 0.09 to 0.25 mEq/L).³ Some transient abnormali­ties were found in infant serum concen­trations of thyroid-stimulating hormone (TSH), blood urea nitrogen, and creatinine; there were no adverse effects on develop­ment. The authors recommend monitoring for TSH abnormalities in infants.

Olanzapine
Olanzapine prescribing information cites a study reporting that 1.8% of the maternal dosage is transferred to breast milk.⁴ Yet the olanzapine prescribing information states, “It is recommended that women receiv­ing olanzapine should not breast-feed.” Olanzapine use during breast-feeding has been studied more than many medications, in part because of a database maintained by the manufacturer. In a study using the man­ufacturer’s database (N = 102) adverse reac­tions were reported in 15.6% of the infants, with the most common being somnolence (3.9%), irritability (2%), tremor (2%), and insomnia (2%).⁵

Other second-generation antipsychotics
Aripiprazole. The only case report of aripip­razole excretion in human breast milk found a concentration of approximately 20% of the maternal plasma level and an M:P ratio of 0.18:0.2.⁶

Asenapine. According to asenapine pre­scribing information⁷ and a literature search, it is not known whether asenapine is excreted in breast milk of humans, although it is found in the milk of lactating rats.

Lurasidone. According to the lurasidone prescribing information⁸ and a literature search, it is not known whether lurasidone is excreted in human breast milk, although it is found in the milk of lactating rats.

Quetiapine. An initial study reported that 0.09% to 0.43% of the maternal dosage of quetiapine was excreted in breast milk.⁹ Further studies found excretion to be 0.09% of maternal dosage, with infant plasma levels reaching 6% of the maternal dos­age.¹⁰ A case series found that one-third of babies exposed to quetiapine during breast-feeding showed some neurodevelopmental delay, although these mothers also were taking other psychotropics.¹¹

Risperidone. A 2000 study¹² of risperidone in lactation reported that 0.84% weight-adjusted maternal risperidone dosage and 3.46% of its metabolite 9-hydroxyrisperi­done is transferred to the infant. A later study showed 2.3% to 4.7% of the mater­nal dosage is transferred, with no adverse events reported in infants.¹³ A case study reported no adverse events and normal neu­rodevelopment in a the child of a mother taking risperidone.¹⁴

Ziprasidone. According to the ziprasidone prescribing information¹⁵ and a literature search, is not known whether ziprasidone is excreted in human breast milk.

See the Table^{4,6-10,12,13,15} for a summary of the evidence levels of second-generation antipsychotics that are excreted in breast milk.

Valproic acid. Because of its high plasma protein binding, valproic acid does not pass readily into the breast milk. Newborns receive approximately 1.4% to 1.7% of the maternal dosage.¹⁶ Caution is advised, however, because of some reported adverse events. One case reported thrombocytopenic purpura and anemia in an infant.¹⁹ Valproic acid is considered to be compatible with breast-feeding with proper monitoring.

Benzodiazepines
Benzodiazepines can be helpful adjunctive medications to aid sleep, which is essen­tial for the mother’s and infant’s health. In a prospectively recruited, retrospec­tively assessed cohort study that evaluated 124 women taking benzodiazepines while breast-feeding, adverse effects, specifically sedation, were noted in 1.6% of infants.²⁰

Future developments in prescribing information
Under a 2008 FDA recommendation, the “Nursing Mothers” section of prescribing information would be replaced with a sec­tion entitled “Lactation.” This new head­ing would include the sub-headings Risk Summary, Clinical Considerations, and Data.¹ It is expected that this new format will be more practical and will help clini­cians and patients make informed deci­sions. The prescribing changes will be in effect on June 30, 2015.²¹

Related Resources
• Massachusetts General Hospital Center for Women’s Mental Health. www.womensmentalhealth.org.
• LactMed. http://toxnet.nlm.nih.gov/newtoxnet/lactmed.htm.
• MOTHERISK. www.motherisk.org/women/ breastfeeding.jsp.

Drug Brand Names
Aripiprazole • Abilify                    Olanzapine • Zyprexa
Asenapine • Saphris                    Quetiapine • Seroquel
Carbamazepine • Tegretol           Risperidone • Risperdal
Lamotrigine • Lamictal                 Valproic acid • Depakene
Lithium • Eskalith, Lithobid           Ziprasidone • Geodon
Lurasidone • Latuda

Disclosure
The author reports no financial relationships with any company whose products are mentioned in this article or with manufacturers of competing products.

HUNTINGTON BEACH, CALIF. – In the opinion of Dr. John M. Oldham, clinicians who deem the alternative personality disorder model of the DSM-5 as too confusing are misguided.

“If you’re going to compare DSM-5 alternative personality disorder model with the DSM-IV model, you have to do a fair comparison,” Dr. Oldham told attendees at the annual meeting of the American College of Psychiatrists. ”In fact, we reduced the number of items that you have to measure by 43%.”

So when people describe the DSM-5’s personality disorders criteria as more complicated, he continued, “what they really mean is, ‘it’s more complicated than what I do,’ not that it’s more complicated than [the] DSM-IV.”

Along with Dr. Andrew E. Skodol, Dr. Oldham cochaired a work group of experts convened by the American Psychiatric Association to update diagnostic criteria related to personality and personality disorders for the DSM-5. “We took our work and our charge seriously,” recalled Dr. Oldham, senior vice president and chief of staff at the Menninger Clinic, Houston. “It was not easy. We had many challenges. A great deal of research has been done in the factor analytic research psychology world around things like the five-factor model of personality. Such terms are not always terribly familiar in clinical medicine, so there was a problem with the lack of familiarity. Then there were vested interests different groups had that were influential in some ways.”

Ultimately, the alternative personality disorder model was placed in section III of the DSM-5. The model enables clinicians “to individually portray the dimensions of the patient’s pathology in a thorough and broad way,” Dr. Oldham explained. “We emphasize impairment in functioning. That’s an important new requirement. So you have to determine, by using the level of functioning scale, whether the person does or doesn’t have moderate or greater impairment. If you have a patient with mild impairment, you can describe what you’re concerned about, but you’re not putting that patient into a diagnostic box of pathology. There is a dimensional scope that enables you to capture many types of patients.”

An empirical study of 337 clinicians demonstrated that in 14 of 18 comparisons, respondents deemed the DSM-5 pathological personality traits as more clinical useful, compared with the DSM-IV, with respect to ease of use, communication of clinical information to other professionals, communication of clinical information to patients, comprehensiveness in describing pathology, and treatment planning (J. Abnorm. Psychol. 2013;122:836-41). “In fact, this was a preference to the new model, which was unfamiliar, compared to the model that these clinicians had been using for 20 years,” Dr. Oldham said.

The study also found that the new DSM-5 personality disorder model was more strongly related to clinical decision making in areas of global functioning, risk assessment, recommended treatment type and intensity, and prognosis.

According to unpublished data from the DSM-5 field trials conducted in the United States and Canada, more than 80% of clinicians in academic and routine clinical practice fields found the new personality disorder criteria “moderately” to “extremely” useful, compared with the DSM-IV. In fact, the respondents rated the new criteria as more useful than other changes to the DSM-5, including those related to bipolar and related disorders, schizophrenia spectrum and other psychotic disorders, and other conditions.

In addition, a test-retest reliability study conducted at 11 academic medical centers found that the new model for borderline personality disorder had a good test-retest reliability (.054), in the same ballpark as that for bipolar I disorder (0.56) and schizophrenia (.50) (Am. J. Psychiatry 2013;170:43-58). “This surprised a lot of people,” Dr. Oldham said.

About 1 year after the DSM-5’s release, Medscape Psychiatry surveyed almost 3,000 clinicians about their impressions of the new guidelines. Of the 2,828 respondents, nearly one-third (28%) were psychiatrists, 22% were psychologists, 13% were family medicine clinicians, and the rest were from other medical fields. The researchers found that 39% of survey respondents were considering the dimensional approaches offered in the new personality disorder criteria of the DSM-5.

“That’s not bad,” Dr. Oldham said.

He reported having no relevant financial conflicts.

[email protected]

On Twitter @dougbrunk

HUNTINGTON BEACH, CALIF. – In the opinion of Dr. John M. Oldham, clinicians who deem the alternative personality disorder model of the DSM-5 as too confusing are misguided.

“If you’re going to compare DSM-5 alternative personality disorder model with the DSM-IV model, you have to do a fair comparison,” Dr. Oldham told attendees at the annual meeting of the American College of Psychiatrists. ”In fact, we reduced the number of items that you have to measure by 43%.”

So when people describe the DSM-5’s personality disorders criteria as more complicated, he continued, “what they really mean is, ‘it’s more complicated than what I do,’ not that it’s more complicated than [the] DSM-IV.”

Along with Dr. Andrew E. Skodol, Dr. Oldham cochaired a work group of experts convened by the American Psychiatric Association to update diagnostic criteria related to personality and personality disorders for the DSM-5. “We took our work and our charge seriously,” recalled Dr. Oldham, senior vice president and chief of staff at the Menninger Clinic, Houston. “It was not easy. We had many challenges. A great deal of research has been done in the factor analytic research psychology world around things like the five-factor model of personality. Such terms are not always terribly familiar in clinical medicine, so there was a problem with the lack of familiarity. Then there were vested interests different groups had that were influential in some ways.”

Ultimately, the alternative personality disorder model was placed in section III of the DSM-5. The model enables clinicians “to individually portray the dimensions of the patient’s pathology in a thorough and broad way,” Dr. Oldham explained. “We emphasize impairment in functioning. That’s an important new requirement. So you have to determine, by using the level of functioning scale, whether the person does or doesn’t have moderate or greater impairment. If you have a patient with mild impairment, you can describe what you’re concerned about, but you’re not putting that patient into a diagnostic box of pathology. There is a dimensional scope that enables you to capture many types of patients.”

An empirical study of 337 clinicians demonstrated that in 14 of 18 comparisons, respondents deemed the DSM-5 pathological personality traits as more clinical useful, compared with the DSM-IV, with respect to ease of use, communication of clinical information to other professionals, communication of clinical information to patients, comprehensiveness in describing pathology, and treatment planning (J. Abnorm. Psychol. 2013;122:836-41). “In fact, this was a preference to the new model, which was unfamiliar, compared to the model that these clinicians had been using for 20 years,” Dr. Oldham said.

The study also found that the new DSM-5 personality disorder model was more strongly related to clinical decision making in areas of global functioning, risk assessment, recommended treatment type and intensity, and prognosis.

According to unpublished data from the DSM-5 field trials conducted in the United States and Canada, more than 80% of clinicians in academic and routine clinical practice fields found the new personality disorder criteria “moderately” to “extremely” useful, compared with the DSM-IV. In fact, the respondents rated the new criteria as more useful than other changes to the DSM-5, including those related to bipolar and related disorders, schizophrenia spectrum and other psychotic disorders, and other conditions.

In addition, a test-retest reliability study conducted at 11 academic medical centers found that the new model for borderline personality disorder had a good test-retest reliability (.054), in the same ballpark as that for bipolar I disorder (0.56) and schizophrenia (.50) (Am. J. Psychiatry 2013;170:43-58). “This surprised a lot of people,” Dr. Oldham said.

About 1 year after the DSM-5’s release, Medscape Psychiatry surveyed almost 3,000 clinicians about their impressions of the new guidelines. Of the 2,828 respondents, nearly one-third (28%) were psychiatrists, 22% were psychologists, 13% were family medicine clinicians, and the rest were from other medical fields. The researchers found that 39% of survey respondents were considering the dimensional approaches offered in the new personality disorder criteria of the DSM-5.

“That’s not bad,” Dr. Oldham said.

He reported having no relevant financial conflicts.

[email protected]

On Twitter @dougbrunk

HUNTINGTON BEACH, CALIF. – In the opinion of Dr. John M. Oldham, clinicians who deem the alternative personality disorder model of the DSM-5 as too confusing are misguided.

“If you’re going to compare DSM-5 alternative personality disorder model with the DSM-IV model, you have to do a fair comparison,” Dr. Oldham told attendees at the annual meeting of the American College of Psychiatrists. ”In fact, we reduced the number of items that you have to measure by 43%.”

So when people describe the DSM-5’s personality disorders criteria as more complicated, he continued, “what they really mean is, ‘it’s more complicated than what I do,’ not that it’s more complicated than [the] DSM-IV.”

Along with Dr. Andrew E. Skodol, Dr. Oldham cochaired a work group of experts convened by the American Psychiatric Association to update diagnostic criteria related to personality and personality disorders for the DSM-5. “We took our work and our charge seriously,” recalled Dr. Oldham, senior vice president and chief of staff at the Menninger Clinic, Houston. “It was not easy. We had many challenges. A great deal of research has been done in the factor analytic research psychology world around things like the five-factor model of personality. Such terms are not always terribly familiar in clinical medicine, so there was a problem with the lack of familiarity. Then there were vested interests different groups had that were influential in some ways.”

Ultimately, the alternative personality disorder model was placed in section III of the DSM-5. The model enables clinicians “to individually portray the dimensions of the patient’s pathology in a thorough and broad way,” Dr. Oldham explained. “We emphasize impairment in functioning. That’s an important new requirement. So you have to determine, by using the level of functioning scale, whether the person does or doesn’t have moderate or greater impairment. If you have a patient with mild impairment, you can describe what you’re concerned about, but you’re not putting that patient into a diagnostic box of pathology. There is a dimensional scope that enables you to capture many types of patients.”

An empirical study of 337 clinicians demonstrated that in 14 of 18 comparisons, respondents deemed the DSM-5 pathological personality traits as more clinical useful, compared with the DSM-IV, with respect to ease of use, communication of clinical information to other professionals, communication of clinical information to patients, comprehensiveness in describing pathology, and treatment planning (J. Abnorm. Psychol. 2013;122:836-41). “In fact, this was a preference to the new model, which was unfamiliar, compared to the model that these clinicians had been using for 20 years,” Dr. Oldham said.

The study also found that the new DSM-5 personality disorder model was more strongly related to clinical decision making in areas of global functioning, risk assessment, recommended treatment type and intensity, and prognosis.

According to unpublished data from the DSM-5 field trials conducted in the United States and Canada, more than 80% of clinicians in academic and routine clinical practice fields found the new personality disorder criteria “moderately” to “extremely” useful, compared with the DSM-IV. In fact, the respondents rated the new criteria as more useful than other changes to the DSM-5, including those related to bipolar and related disorders, schizophrenia spectrum and other psychotic disorders, and other conditions.

In addition, a test-retest reliability study conducted at 11 academic medical centers found that the new model for borderline personality disorder had a good test-retest reliability (.054), in the same ballpark as that for bipolar I disorder (0.56) and schizophrenia (.50) (Am. J. Psychiatry 2013;170:43-58). “This surprised a lot of people,” Dr. Oldham said.

About 1 year after the DSM-5’s release, Medscape Psychiatry surveyed almost 3,000 clinicians about their impressions of the new guidelines. Of the 2,828 respondents, nearly one-third (28%) were psychiatrists, 22% were psychologists, 13% were family medicine clinicians, and the rest were from other medical fields. The researchers found that 39% of survey respondents were considering the dimensional approaches offered in the new personality disorder criteria of the DSM-5.

“That’s not bad,” Dr. Oldham said.

He reported having no relevant financial conflicts.

[email protected]

On Twitter @dougbrunk

CASE Delusional and aggressive
Mr. P, age 78, of Filipino heritage, is brought to the psychiatric hospital because he has been verbally aggressive toward his wife for sev­eral weeks. He has no history of a psychiatric diagnosis or inpatient psychiatric hospitaliza­tion, and no history of taking any psychotropic medications.

According to his wife, Mr. P has been rumi­nating about his father, who died in World War II, saying that “the Japanese never gave his body back” to him. Also, his wife describes 3 weeks of physically aggressive behavior, such as throwing punches; the last episode was 2 days before admission.

Mr. P is not bathing, eating, taking his medi­cations, and attending to his activities of daily living. He sleeps for only 1 to 2 hours a night; is irritable and easily distractible; and experi­ences flight of ideas. Mr. P has been buying lottery tickets, telling his daughter that he will become a millionaire and then buy a house in the Philippines.

Mr. P reports depressed mood, but no other depressive symptoms are present. He reports no suicidal or homicidal ideations, auditory or visual hallucinations, or anxiety symptoms. He has no history of substance abuse.

What diagnosis would you give Mr. P?
   a) late-onset bipolar disorder
   b) Alzheimer’s disease
   c) major depressive disorder
   d) frontotemporal dementia

The authors’ observations
Bipolar disorder in later life is a complex and confounding neuropsychiatric syn­drome with diagnostic and therapeutic challenges. The disorder can affect people of all ages and is not uncommon among geriatric patients, with a 1-year prevalence in United States of 0.4%.¹ In one study, 10% of new bipolar disorder cases were found to occur after age 50.² As the American population grows older, the number of bipolar disorder cases among seniors is expected to increase.³

It was once thought that symptoms of bipolar disorder disappear with age; newer research has disproved this theory, and proposes that untreated bipolar dis­order worsens over time.⁴ Persons who are given the diagnosis later in life could have had bipolar disorder for decades, but symptoms became more noticeable and problematic with age.⁵

Common symptoms in geriatric patients can differ from what we might expect in younger patients: agitation, hyperactivity, irritability, confusion, and psychosis.⁶ When the disorder presents in patients age >60, it can be severe, with significant changes in cognitive function, including difficulties with memory, perception, judgment, and problem-solving.^7,8

HISTORY Medical comorbidities
Mr. P emigrated from the Philippines 20 years ago, is married, and lives with his wife. He has 3 brothers; his parents were divorced, and his mother remarried. Mr. P completed high school.

Mr. P has an extensive medical history: diabetes mellitus, hypertension, dyslipid­emia, and recent double coronary artery bypass grafting. He is taking several medi­cations: sitagliptin, 25 mg/d; pantoprazole, 5 mg/d; metformin, 1,000 mg/d; rivaroxaban, 20 mg/d; amiodarone, 200 mg/d; metoprolol, 12.5 mg/d; olmesartan medoxomil, 40 mg/d; aspirin, 81 mg/d; simvastatin, 10 mg/d; eszopi­clone, 3 mg at bedtime; and amlodipine, 5 mg at bedtime.

Mr. P was following up with his primary care physician for his medical conditions and was adherent with treatment until 1 week before he was admitted to our facility.

The authors’ observations
Always rule out medical causes in a case of new-onset mania, which is particu­larly important in geriatric patients. Older patients with new-onset mania are more than twice as likely to have a comorbid neurologic disorder.⁹ Neurologic causes of late-onset mania include:
   • stroke
   • tumor
   • epilepsy
   • Huntington’s disease and other movement disorders
   • multiple sclerosis and other white-matter diseases
   • head trauma
   • infection (such as neurosyphilis)
   • Creutzfeldt-Jakob disease
   • frontotemporal dementia.¹⁰

Mr. P’s presentation of psychomotor agitation, impaired functioning, decreased need for sleep, increased energy, hyperver­bal speech, and complex paranoid delu­sions meets DSM-5 criteria for bipolar disorder, manic phase. In addition, older manic patients frequently present with confusion, disorientation, and distract­ibility. Younger patients with mania often present with euphoric moods and gran­diosity; in contrast, geriatric patients are more likely to show a mixture of depressed affect and manic symptoms (pressured speech and a decreased need for sleep).^11-15

We considered an emerging neurode­generative process, because dementia can present early with disinhibition, lability, and other behavioral disturbances, includ­ing classic manic syndromes.¹⁶ Although we could not fully rule out a neurode­generative process in the initial phase of treatment, Mr. P’s longitudinal course demonstrated no change in baseline cog­nitive function and no evidence of subse­quent decline, making dementia unlikely.¹⁷

Patients with frontotemporal demen­tia are more likely to present initially to a psychiatrist than to a neurologist.¹⁸

Frontotemporal dementia is a progressive neurodegenerative disease that affects the frontal and temporal cortices; it is a com­mon cause of dementia in patients age <65.¹⁹ Frontotemporal dementia is char­acterized by insidious behavioral and personality changes; often, the initial pre­sentation lacks any clear neurologic signs or symptoms. Key features include apa­thy, disinhibition, loss of sympathy and empathy, repetitive motor behaviors, and overeating.²⁰

Mr. P’s symptoms stabilized with dival­proex sprinkles and risperidone. There was no evidence of decline in memory, social interaction, or behavior.

EVALUATION Paranoia
On mental status exam, Mr. P has an appropri­ate appearance; he is clean and shaven, with good eye contact. Muscular tone and gait are within normal limits. Level of activity is increased; he exhibits psychomotor agitation. Speech is rapid, over-productive, and loud; thought process shows flight of ideas, and thought associations are circumstantial.

Mr. P has paranoid delusions about the staff trying to hurt him. His judgment is poor, evidenced by an inability to take care of him­self. Insight is minimal, as seen by noncompli­ance with treatment. Mr. P is oriented only to person and place. His mood is anxious; affect is labile.

Complete blood count, comprehensive met­abolic profile, blood alcohol level, urine analy­sis, urine toxicology, electrocardiogram, and CT scan of the head are within normal limits.

Mr. P is given a diagnosis of mood disorder due to general medical condition, psychotic disorder due to general medical condition. The team rules out acute delirium, bipolar I disor­der, and neurodegenerative disorders such as frontotemporal dementia.

Mr. P is maintained on pre-admission medi­cations for his medical conditions. A mood sta­bilizer, divalproex sprinkles, 250 mg/d, is added.

Once on the unit, Mr. P is re-evaluated. Divalproex is increased to 500 mg/d; risperi­done, 0.5 mg/d, is added to address paranoia. Mr. P also receives group and individual psy­chotherapy. He does not participate in neuro­psychological testing, and no single-photon emission CT analysis is done. Mr. P remains in the hospital for 2 weeks. After a family meeting, his daughter says she feels comfortable taking Mr. P home. He follows up in the outpatient clinic and is doing well.

The authors’ observations
Treating geriatric patients with bipolar disorder requires attention to several fac­tors (Table). Older patients might tolerate or metabolize medications differently than younger adults, and therefore may need a different dosage. Older patients are more likely to have comorbid medical conditions and to be taking medications for those ail­ments. Treatment is much more compli­cated for this age group because physicians need to account for possible drug-drug interactions.²¹

Psychotherapy can be a valuable addition to pharmacotherapy in older adults. Some psychotherapy programs are specifically geared to older bipolar disorder patients.^22,23

Use of divalproex sodium in older patients
First, perform baseline laboratory tests: complete blood count, liver function, and electrocardiogram. Initiate divalproex sodium, 250 mg at bedtime, increasing the dosage every 3 to 5 days by 250 mg, with a target dose of 500 to 2,000 mg/d (divided into 2 or 3 doses). Monitor serum levels; levels of 29 to 100 μg/mL are effective and well tolerated. Common side effects include excess sedation, ataxia, tremor, nausea, and, rarely, hepatotoxicity, leuko­penia, and thrombocytopenia.²⁴

Use of lithium in geriatric patients
First, perform baseline laboratory tests: electrolytes, creatinine, blood urea nitro­gen, urine, thyroid stimulating hormone, and electrocardiogram. Starting dosage is 300 mg at bedtime (150 mg for frail cachec­tic patients). Monitor serum levels 12 hours after last dose, adjusting dosage every 5 days until a target serum level of 0.5 to 0.8 mEq/L is reached. Common dosages for geriatric patients are 300 to 600 mg/d, which often can be given as a single bed­time dose. Cautions: When using lithium with a thiazide diuretic or nonsteroidal anti-inflammatory drug, watch for dehy­dration, vomiting, and diarrhea, which will elevate the serum lithium level. Side effects include ataxia, tremor, urinary frequency, thirst, nausea, diarrhea, hypothyroidism, and exacerbation of psoriasis. Once sta­bilized, monitor the serum lithium level, thyroid-stimulating hormone, and kidney function every 3 to 6 months.²⁴

Bottom Line
In geriatric patients, bipolar disorder can present with agitation, irritability, confusion, and psychosis, rather than euphoric mood and grandiosity. When you suspect bipolar disorder in an older patient, first rule out medical causes of symptoms. When selecting treatment, consider comorbid medical conditions and possible drug-drug interactions.

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

CASE Delusional and aggressive
Mr. P, age 78, of Filipino heritage, is brought to the psychiatric hospital because he has been verbally aggressive toward his wife for sev­eral weeks. He has no history of a psychiatric diagnosis or inpatient psychiatric hospitaliza­tion, and no history of taking any psychotropic medications.

According to his wife, Mr. P has been rumi­nating about his father, who died in World War II, saying that “the Japanese never gave his body back” to him. Also, his wife describes 3 weeks of physically aggressive behavior, such as throwing punches; the last episode was 2 days before admission.

Mr. P is not bathing, eating, taking his medi­cations, and attending to his activities of daily living. He sleeps for only 1 to 2 hours a night; is irritable and easily distractible; and experi­ences flight of ideas. Mr. P has been buying lottery tickets, telling his daughter that he will become a millionaire and then buy a house in the Philippines.

Mr. P reports depressed mood, but no other depressive symptoms are present. He reports no suicidal or homicidal ideations, auditory or visual hallucinations, or anxiety symptoms. He has no history of substance abuse.

What diagnosis would you give Mr. P?
   a) late-onset bipolar disorder
   b) Alzheimer’s disease
   c) major depressive disorder
   d) frontotemporal dementia

The authors’ observations
Bipolar disorder in later life is a complex and confounding neuropsychiatric syn­drome with diagnostic and therapeutic challenges. The disorder can affect people of all ages and is not uncommon among geriatric patients, with a 1-year prevalence in United States of 0.4%.¹ In one study, 10% of new bipolar disorder cases were found to occur after age 50.² As the American population grows older, the number of bipolar disorder cases among seniors is expected to increase.³

It was once thought that symptoms of bipolar disorder disappear with age; newer research has disproved this theory, and proposes that untreated bipolar dis­order worsens over time.⁴ Persons who are given the diagnosis later in life could have had bipolar disorder for decades, but symptoms became more noticeable and problematic with age.⁵

Common symptoms in geriatric patients can differ from what we might expect in younger patients: agitation, hyperactivity, irritability, confusion, and psychosis.⁶ When the disorder presents in patients age >60, it can be severe, with significant changes in cognitive function, including difficulties with memory, perception, judgment, and problem-solving.^7,8

HISTORY Medical comorbidities
Mr. P emigrated from the Philippines 20 years ago, is married, and lives with his wife. He has 3 brothers; his parents were divorced, and his mother remarried. Mr. P completed high school.

Mr. P has an extensive medical history: diabetes mellitus, hypertension, dyslipid­emia, and recent double coronary artery bypass grafting. He is taking several medi­cations: sitagliptin, 25 mg/d; pantoprazole, 5 mg/d; metformin, 1,000 mg/d; rivaroxaban, 20 mg/d; amiodarone, 200 mg/d; metoprolol, 12.5 mg/d; olmesartan medoxomil, 40 mg/d; aspirin, 81 mg/d; simvastatin, 10 mg/d; eszopi­clone, 3 mg at bedtime; and amlodipine, 5 mg at bedtime.

Mr. P was following up with his primary care physician for his medical conditions and was adherent with treatment until 1 week before he was admitted to our facility.

The authors’ observations
Always rule out medical causes in a case of new-onset mania, which is particu­larly important in geriatric patients. Older patients with new-onset mania are more than twice as likely to have a comorbid neurologic disorder.⁹ Neurologic causes of late-onset mania include:
   • stroke
   • tumor
   • epilepsy
   • Huntington’s disease and other movement disorders
   • multiple sclerosis and other white-matter diseases
   • head trauma
   • infection (such as neurosyphilis)
   • Creutzfeldt-Jakob disease
   • frontotemporal dementia.¹⁰

Mr. P’s presentation of psychomotor agitation, impaired functioning, decreased need for sleep, increased energy, hyperver­bal speech, and complex paranoid delu­sions meets DSM-5 criteria for bipolar disorder, manic phase. In addition, older manic patients frequently present with confusion, disorientation, and distract­ibility. Younger patients with mania often present with euphoric moods and gran­diosity; in contrast, geriatric patients are more likely to show a mixture of depressed affect and manic symptoms (pressured speech and a decreased need for sleep).^11-15

We considered an emerging neurode­generative process, because dementia can present early with disinhibition, lability, and other behavioral disturbances, includ­ing classic manic syndromes.¹⁶ Although we could not fully rule out a neurode­generative process in the initial phase of treatment, Mr. P’s longitudinal course demonstrated no change in baseline cog­nitive function and no evidence of subse­quent decline, making dementia unlikely.¹⁷

Patients with frontotemporal demen­tia are more likely to present initially to a psychiatrist than to a neurologist.¹⁸

Frontotemporal dementia is a progressive neurodegenerative disease that affects the frontal and temporal cortices; it is a com­mon cause of dementia in patients age <65.¹⁹ Frontotemporal dementia is char­acterized by insidious behavioral and personality changes; often, the initial pre­sentation lacks any clear neurologic signs or symptoms. Key features include apa­thy, disinhibition, loss of sympathy and empathy, repetitive motor behaviors, and overeating.²⁰

Mr. P’s symptoms stabilized with dival­proex sprinkles and risperidone. There was no evidence of decline in memory, social interaction, or behavior.

EVALUATION Paranoia
On mental status exam, Mr. P has an appropri­ate appearance; he is clean and shaven, with good eye contact. Muscular tone and gait are within normal limits. Level of activity is increased; he exhibits psychomotor agitation. Speech is rapid, over-productive, and loud; thought process shows flight of ideas, and thought associations are circumstantial.

Mr. P has paranoid delusions about the staff trying to hurt him. His judgment is poor, evidenced by an inability to take care of him­self. Insight is minimal, as seen by noncompli­ance with treatment. Mr. P is oriented only to person and place. His mood is anxious; affect is labile.

Complete blood count, comprehensive met­abolic profile, blood alcohol level, urine analy­sis, urine toxicology, electrocardiogram, and CT scan of the head are within normal limits.

Mr. P is given a diagnosis of mood disorder due to general medical condition, psychotic disorder due to general medical condition. The team rules out acute delirium, bipolar I disor­der, and neurodegenerative disorders such as frontotemporal dementia.

Mr. P is maintained on pre-admission medi­cations for his medical conditions. A mood sta­bilizer, divalproex sprinkles, 250 mg/d, is added.

Once on the unit, Mr. P is re-evaluated. Divalproex is increased to 500 mg/d; risperi­done, 0.5 mg/d, is added to address paranoia. Mr. P also receives group and individual psy­chotherapy. He does not participate in neuro­psychological testing, and no single-photon emission CT analysis is done. Mr. P remains in the hospital for 2 weeks. After a family meeting, his daughter says she feels comfortable taking Mr. P home. He follows up in the outpatient clinic and is doing well.

The authors’ observations
Treating geriatric patients with bipolar disorder requires attention to several fac­tors (Table). Older patients might tolerate or metabolize medications differently than younger adults, and therefore may need a different dosage. Older patients are more likely to have comorbid medical conditions and to be taking medications for those ail­ments. Treatment is much more compli­cated for this age group because physicians need to account for possible drug-drug interactions.²¹

Psychotherapy can be a valuable addition to pharmacotherapy in older adults. Some psychotherapy programs are specifically geared to older bipolar disorder patients.^22,23

Use of divalproex sodium in older patients
First, perform baseline laboratory tests: complete blood count, liver function, and electrocardiogram. Initiate divalproex sodium, 250 mg at bedtime, increasing the dosage every 3 to 5 days by 250 mg, with a target dose of 500 to 2,000 mg/d (divided into 2 or 3 doses). Monitor serum levels; levels of 29 to 100 μg/mL are effective and well tolerated. Common side effects include excess sedation, ataxia, tremor, nausea, and, rarely, hepatotoxicity, leuko­penia, and thrombocytopenia.²⁴

Use of lithium in geriatric patients
First, perform baseline laboratory tests: electrolytes, creatinine, blood urea nitro­gen, urine, thyroid stimulating hormone, and electrocardiogram. Starting dosage is 300 mg at bedtime (150 mg for frail cachec­tic patients). Monitor serum levels 12 hours after last dose, adjusting dosage every 5 days until a target serum level of 0.5 to 0.8 mEq/L is reached. Common dosages for geriatric patients are 300 to 600 mg/d, which often can be given as a single bed­time dose. Cautions: When using lithium with a thiazide diuretic or nonsteroidal anti-inflammatory drug, watch for dehy­dration, vomiting, and diarrhea, which will elevate the serum lithium level. Side effects include ataxia, tremor, urinary frequency, thirst, nausea, diarrhea, hypothyroidism, and exacerbation of psoriasis. Once sta­bilized, monitor the serum lithium level, thyroid-stimulating hormone, and kidney function every 3 to 6 months.²⁴

Bottom Line
In geriatric patients, bipolar disorder can present with agitation, irritability, confusion, and psychosis, rather than euphoric mood and grandiosity. When you suspect bipolar disorder in an older patient, first rule out medical causes of symptoms. When selecting treatment, consider comorbid medical conditions and possible drug-drug interactions.

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

CASE Delusional and aggressive
Mr. P, age 78, of Filipino heritage, is brought to the psychiatric hospital because he has been verbally aggressive toward his wife for sev­eral weeks. He has no history of a psychiatric diagnosis or inpatient psychiatric hospitaliza­tion, and no history of taking any psychotropic medications.

According to his wife, Mr. P has been rumi­nating about his father, who died in World War II, saying that “the Japanese never gave his body back” to him. Also, his wife describes 3 weeks of physically aggressive behavior, such as throwing punches; the last episode was 2 days before admission.

Mr. P is not bathing, eating, taking his medi­cations, and attending to his activities of daily living. He sleeps for only 1 to 2 hours a night; is irritable and easily distractible; and experi­ences flight of ideas. Mr. P has been buying lottery tickets, telling his daughter that he will become a millionaire and then buy a house in the Philippines.

Mr. P reports depressed mood, but no other depressive symptoms are present. He reports no suicidal or homicidal ideations, auditory or visual hallucinations, or anxiety symptoms. He has no history of substance abuse.

What diagnosis would you give Mr. P?
   a) late-onset bipolar disorder
   b) Alzheimer’s disease
   c) major depressive disorder
   d) frontotemporal dementia

The authors’ observations
Bipolar disorder in later life is a complex and confounding neuropsychiatric syn­drome with diagnostic and therapeutic challenges. The disorder can affect people of all ages and is not uncommon among geriatric patients, with a 1-year prevalence in United States of 0.4%.¹ In one study, 10% of new bipolar disorder cases were found to occur after age 50.² As the American population grows older, the number of bipolar disorder cases among seniors is expected to increase.³

It was once thought that symptoms of bipolar disorder disappear with age; newer research has disproved this theory, and proposes that untreated bipolar dis­order worsens over time.⁴ Persons who are given the diagnosis later in life could have had bipolar disorder for decades, but symptoms became more noticeable and problematic with age.⁵

Common symptoms in geriatric patients can differ from what we might expect in younger patients: agitation, hyperactivity, irritability, confusion, and psychosis.⁶ When the disorder presents in patients age >60, it can be severe, with significant changes in cognitive function, including difficulties with memory, perception, judgment, and problem-solving.^7,8

HISTORY Medical comorbidities
Mr. P emigrated from the Philippines 20 years ago, is married, and lives with his wife. He has 3 brothers; his parents were divorced, and his mother remarried. Mr. P completed high school.

Mr. P has an extensive medical history: diabetes mellitus, hypertension, dyslipid­emia, and recent double coronary artery bypass grafting. He is taking several medi­cations: sitagliptin, 25 mg/d; pantoprazole, 5 mg/d; metformin, 1,000 mg/d; rivaroxaban, 20 mg/d; amiodarone, 200 mg/d; metoprolol, 12.5 mg/d; olmesartan medoxomil, 40 mg/d; aspirin, 81 mg/d; simvastatin, 10 mg/d; eszopi­clone, 3 mg at bedtime; and amlodipine, 5 mg at bedtime.

Mr. P was following up with his primary care physician for his medical conditions and was adherent with treatment until 1 week before he was admitted to our facility.

The authors’ observations
Always rule out medical causes in a case of new-onset mania, which is particu­larly important in geriatric patients. Older patients with new-onset mania are more than twice as likely to have a comorbid neurologic disorder.⁹ Neurologic causes of late-onset mania include:
   • stroke
   • tumor
   • epilepsy
   • Huntington’s disease and other movement disorders
   • multiple sclerosis and other white-matter diseases
   • head trauma
   • infection (such as neurosyphilis)
   • Creutzfeldt-Jakob disease
   • frontotemporal dementia.¹⁰

Mr. P’s presentation of psychomotor agitation, impaired functioning, decreased need for sleep, increased energy, hyperver­bal speech, and complex paranoid delu­sions meets DSM-5 criteria for bipolar disorder, manic phase. In addition, older manic patients frequently present with confusion, disorientation, and distract­ibility. Younger patients with mania often present with euphoric moods and gran­diosity; in contrast, geriatric patients are more likely to show a mixture of depressed affect and manic symptoms (pressured speech and a decreased need for sleep).^11-15

We considered an emerging neurode­generative process, because dementia can present early with disinhibition, lability, and other behavioral disturbances, includ­ing classic manic syndromes.¹⁶ Although we could not fully rule out a neurode­generative process in the initial phase of treatment, Mr. P’s longitudinal course demonstrated no change in baseline cog­nitive function and no evidence of subse­quent decline, making dementia unlikely.¹⁷

Patients with frontotemporal demen­tia are more likely to present initially to a psychiatrist than to a neurologist.¹⁸

Frontotemporal dementia is a progressive neurodegenerative disease that affects the frontal and temporal cortices; it is a com­mon cause of dementia in patients age <65.¹⁹ Frontotemporal dementia is char­acterized by insidious behavioral and personality changes; often, the initial pre­sentation lacks any clear neurologic signs or symptoms. Key features include apa­thy, disinhibition, loss of sympathy and empathy, repetitive motor behaviors, and overeating.²⁰

Mr. P’s symptoms stabilized with dival­proex sprinkles and risperidone. There was no evidence of decline in memory, social interaction, or behavior.

EVALUATION Paranoia
On mental status exam, Mr. P has an appropri­ate appearance; he is clean and shaven, with good eye contact. Muscular tone and gait are within normal limits. Level of activity is increased; he exhibits psychomotor agitation. Speech is rapid, over-productive, and loud; thought process shows flight of ideas, and thought associations are circumstantial.

Mr. P has paranoid delusions about the staff trying to hurt him. His judgment is poor, evidenced by an inability to take care of him­self. Insight is minimal, as seen by noncompli­ance with treatment. Mr. P is oriented only to person and place. His mood is anxious; affect is labile.

Complete blood count, comprehensive met­abolic profile, blood alcohol level, urine analy­sis, urine toxicology, electrocardiogram, and CT scan of the head are within normal limits.

Mr. P is given a diagnosis of mood disorder due to general medical condition, psychotic disorder due to general medical condition. The team rules out acute delirium, bipolar I disor­der, and neurodegenerative disorders such as frontotemporal dementia.

Mr. P is maintained on pre-admission medi­cations for his medical conditions. A mood sta­bilizer, divalproex sprinkles, 250 mg/d, is added.

Once on the unit, Mr. P is re-evaluated. Divalproex is increased to 500 mg/d; risperi­done, 0.5 mg/d, is added to address paranoia. Mr. P also receives group and individual psy­chotherapy. He does not participate in neuro­psychological testing, and no single-photon emission CT analysis is done. Mr. P remains in the hospital for 2 weeks. After a family meeting, his daughter says she feels comfortable taking Mr. P home. He follows up in the outpatient clinic and is doing well.

The authors’ observations
Treating geriatric patients with bipolar disorder requires attention to several fac­tors (Table). Older patients might tolerate or metabolize medications differently than younger adults, and therefore may need a different dosage. Older patients are more likely to have comorbid medical conditions and to be taking medications for those ail­ments. Treatment is much more compli­cated for this age group because physicians need to account for possible drug-drug interactions.²¹

Psychotherapy can be a valuable addition to pharmacotherapy in older adults. Some psychotherapy programs are specifically geared to older bipolar disorder patients.^22,23

Use of divalproex sodium in older patients
First, perform baseline laboratory tests: complete blood count, liver function, and electrocardiogram. Initiate divalproex sodium, 250 mg at bedtime, increasing the dosage every 3 to 5 days by 250 mg, with a target dose of 500 to 2,000 mg/d (divided into 2 or 3 doses). Monitor serum levels; levels of 29 to 100 μg/mL are effective and well tolerated. Common side effects include excess sedation, ataxia, tremor, nausea, and, rarely, hepatotoxicity, leuko­penia, and thrombocytopenia.²⁴

Use of lithium in geriatric patients
First, perform baseline laboratory tests: electrolytes, creatinine, blood urea nitro­gen, urine, thyroid stimulating hormone, and electrocardiogram. Starting dosage is 300 mg at bedtime (150 mg for frail cachec­tic patients). Monitor serum levels 12 hours after last dose, adjusting dosage every 5 days until a target serum level of 0.5 to 0.8 mEq/L is reached. Common dosages for geriatric patients are 300 to 600 mg/d, which often can be given as a single bed­time dose. Cautions: When using lithium with a thiazide diuretic or nonsteroidal anti-inflammatory drug, watch for dehy­dration, vomiting, and diarrhea, which will elevate the serum lithium level. Side effects include ataxia, tremor, urinary frequency, thirst, nausea, diarrhea, hypothyroidism, and exacerbation of psoriasis. Once sta­bilized, monitor the serum lithium level, thyroid-stimulating hormone, and kidney function every 3 to 6 months.²⁴

Bottom Line
In geriatric patients, bipolar disorder can present with agitation, irritability, confusion, and psychosis, rather than euphoric mood and grandiosity. When you suspect bipolar disorder in an older patient, first rule out medical causes of symptoms. When selecting treatment, consider comorbid medical conditions and possible drug-drug interactions.

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

CASE Broken down
Ms. E, age 20, is a college student who has had major depressive disorder for several years and a genetic bone disease (osteogenesis imperfecta, mixed type III and IV). She pres­ents with depression, anxiety, and suicidal ideation. She reports recent worsening of her depressive symptoms, including anhedonia, excessive sleep, difficulty concentrating, and feeling overwhelmed, hopeless, and worth­less. She also describes frequent thoughts of suicide with the plan of putting herself in oncoming traffic, although she has no history of suicide attempts.

Previously, her primary care physician prescribed lorazepam, 0.5 mg, as needed for anxiety, and sertraline, 100 mg/d, for depres­sion and anxiety. She experienced only partial improvement in symptoms, however.

In addition to depressive symptoms, Ms. E describes manic symptoms lasting for as long as 3 to 5 days, including decreased need for sleep, increased energy, pressured speech, racing thoughts, distractibility, spending excessive money on cosmetics, and risking her safety—given her skeletal disorder— by participating in high-impact stage-combat classes. She denies auditory and visual hallucinations, homicidal ideation, and delusions.

The medical history is significant for osteogenesis imperfecta, which has caused 62 fractures and required 16 surgeries. Ms. E is a theater major who, despite her short stature and wheelchair use, reports enjoying her acting career and says she does not feel demoralized by her medical condition. She describes overcoming her physical disabili­ties with pride and confidence. However, her recent worsening mood symptoms have left her unable to concentrate and feeling over­whelmed with school.

Ms. E is voluntarily admitted to an inpatient psychiatric unit with a diagnosis of bipolar I disorder with rapid cycling, most recent epi­sode mixed. Because of her bone fragility, the treatment team considers what would be an appropriate course of drug treatment to con­trol bipolar symptoms while minimizing risk of bone loss.

Which medications are associated with decreased bone mineral density?
   a) citalopram
   b) haloperidol
   c) carbamazepine
   d) paliperidone
   e) all of the above

The authors’ observations
Osteogenesis imperfecta is a genetic condi­tion caused by mutations in genes impli­cated in collagen production. As a result, bones are brittle and prone to fracture. Different classes of psychotropics have been shown to increase risk of bone frac­tures through a variety of mechanisms. Clinicians often must choose appropri­ate pharmacotherapy for patients at high risk of fracture, including postmenopausal women, older patients, malnourished per­sons, and those with hormonal deficien­cies leading to osteoporosis.

To assist our clinical decision-making, we reviewed the literature to establish appropriate management of a patient with increased bone fragility and new-onset bipolar disorder. We considered all classes of medications used to treat bipolar disor­der, including antipsychotics, antidepres­sants, lithium, and anticonvulsants.

Antipsychotics
In population-based studies, prolactin-elevating antipsychotics have been associ­ated with decreased bone mineral density and increased risk of fracture.¹ Additional studies on geriatric and non-geriatric popu­lations have supported these findings.^2,3

The mechanism through which fracture risk is increased likely is related to antipsy­chotics’ effect on serum prolactin and corti­sol levels. Antipsychotics act as antagonists on D2 receptors in the hypothalamic tubero-infundibular pathway, therefore preventing inhibition of prolactin. Long-term elevation in serum prolactin can cause loss of bone mineral density through secondary hypo­gonadism and direct effects on target tis­sues. Additional modifying factors include smoking and estrogen use.

The degree to which antipsychotics increase fracture risk might be related to the degree of serum prolactin elevation.⁴ Antipsychotics previously have been grouped by the degree of prolactin eleva­tion, categorizing them as high, medium, and low or no potential to elevate serum prolactin.⁴ Based on this classification, typical antipsychotics, risperidone, and paliperidone have the highest potential to elevate prolactin. Accordingly, antipsychot­ics with the lowest fracture risk are those that have the lowest risk of serum prolactin elevation: ziprasidone, asenapine, quetiap­ine, and clozapine. Aripiprazole may lower prolactin in some patients. This is sup­ported by studies noting reduced bone min­eral density^5,6 and increased risk of fracture¹ with high-potential vs low- or no-potential antipsychotics. Because of these findings, it is crucial to consider the potential risk of prolactin elevation when treating patients at increased risk of fracture. Providers should consider low/no potential antipsy­chotic medications before considering those with medium or high potential (Table).

Antidepressants
In a meta-analysis, antidepressants were shown to increase fracture risk by 70% to 90%.² However, the relative risk varies by antidepressant class. Several studies have shown that selective serotonin reuptake inhibitors (SSRIs) are associated with a higher risk of fracture compared with tri­cyclic antidepressants (TCAs).⁷ In addition, antidepressants with a high affinity for the serotonin transporter, including citalopram, fluoxetine, fluvoxamine, paroxetine, sertra­line, and imipramine, have been associated with greater risk of osteoporotic fracture compared with those with low affinity.⁸

The mechanisms by which antidepres­sants increase fracture risk are complex, although the strongest evidence implicates a direct effect on bone metabolism via the 5-HTT receptor. This receptor, found on osteoblasts and osteoclasts, plays an impor­tant role in bone metabolism; it is through this receptor that SSRIs might inhibit osteo­blasts and promote osteoclast activity, thereby disrupting bone microarchitecture. Additional studies are needed to further describe the mechanism of the association among antidepressants, bone mineral den­sity, and fracture risk.

Fracture risk is associated with dura­tion of use rather than dosage. Population-based studies show a higher fracture risk for new users of TCAs compared with continuous users, and the risk of fracture with SSRIs seems to increase slightly over time.⁹ No association has been identified between fracture risk and antidepressant dosage. According to the literature, drugs with low affinity for the serotonin trans­porter, such as maprotiline and mirtazap­ine, likely are the safest antidepressants for patients at increased risk of fracture. Options also include other TCAs and any antidepressant with low affinity for the serotonin receptor.^7,8 

Lithium
Studies on lithium and bone mineral den­sity have shown mixed results. Older stud­ies found that lithium had a negative or no effect on bone mineral density or the para­thyroid hormone level.¹⁰ More recent inves­tigations, however, suggest that the drug has a protective effect on bone mineral den­sity, although this has not been replicated in all studies.

In a mouse model, lithium has been shown to enhance bone formation and improve bone mass, at least in part by activation of the Wnt signaling pathway through an inhibitory effect on glycogen synthase kinase-3β.¹¹ In humans, lithium-treated adults had lower serum alkaline phosphate, osteocalcin, and C-telopeptide levels compared with controls, suggesting a state of decreased bone remodeling and increased turnover.¹² There is a paucity of clinical data on the effect of lithium on frac­ture risk. Additional studies are necessary to elucidate lithium’s mechanism on bone mineral density and determine the magni­tude of the clinical effect.

Anticonvulsants
The association among anticonvulsants, decreased bone mineral density, and increased risk of fracture is well-established in the literature.¹³ However, causality is dif­ficult to determine, because many studies were of patients with a seizure disorder, who often have additional risk factors for fracture, including seizure-related trauma, drowsiness, and slowed reflexes.

Mechanisms through which anticon­vulsants increase fracture risk include increased bone resorption, secondary hypo­parathyroidism, and pseudohypoparathy­roidism. Markers of bone resorption were elevated in patients receiving an antiepi­leptic.¹⁴ This effect might be enhanced by co-administration of cytochrome P450 (CYP450) enzyme-inducing anticon­vulsants and CYP450 enzyme-inhibiting medications, such as valproate. Long-term treatment with valproate may produce reduction of bone mass and increased risk of fractures; however, other studies dis­agree with this finding.¹⁵

In addition to CYP450-inducing effects, phenytoin, carbamezapine, and phenobar­bital can increase catabolism of vitamin D, which is associated with osteomalacia.¹⁴ This results in decreased intestinal absorp­tion of calcium, hypocalcemia, and sec­ondary hyperparathyroidism, which also increases fracture risk. Anticonvulsants also might increase resistance to pseudo­hypoparathyroidism and inhibit calcitonin secretion.

Lamotrigine has not been shown to interfere with bone accrual¹⁶ and may be a safer mood stabilizer for patients at high risk of fracture. For patients at increased risk of fracture, it is important to select an anticonvulsant wisely to minimize frac­ture risk.

How would you treat Ms. E during her hospitalization for bipolar disorder?
   a) carbamazepine
   b) lithium
   c) risperidone
   d) mirtazapine

TREATMENT Minimizing polypharmacy
Because many pharmacotherapeutic options for managing bipolar disorder can increase the risk of fracture, clinicians must be aware of the relative risk of each class of medication and each individual drug. We initiated lithium, 300 mg, 3 times a day, to stabilize Ms. E’s mood. Although clinical data are inconclusive regarding lithium’s effect on fracture risk, we felt that the benefit of acute mood stabiliza­tion outweighed the risk of decreased bone mineral index.

We selected aripiprazole, 10 mg/d, as an adjunctive treatment because of its minimal effect on serum prolactin levels.⁴ We con­sidered prescribing an antidepressant but decided against it because we were concerned about manic switching.

Polypharmacy is another important con­sideration for Ms. E. Several studies have identified polypharmacy, particularly with antipsychotics, as an independent risk factor for fracture.³ Therefore, we sought to minimize the number of medications Ms. E receives. Although lithium monotherapy is an option, we thought that her mood symptoms were severe enough that the risk of inadequately treating her bipolar symptoms outweighed the additional risk of fracture from dual ther­apy with lithium and aripiprazole. Untreated or inadequately treated depression is associ­ated with a higher fracture risk. Therefore, we avoided prescribing >2 medications to mitigate any excessive risk of fracture from polypharmacy.

Bottom Line
Different classes of medications—antipsychotics, anticonvulsants, antidepressants, and lithium—used for treating bipolar disorder have been shown to increase risk of bone fracture through a variety of mechanisms. Anticonvulsants and prolactin-elevating antipsychotics are associated with increased fracture risk; evidence on lithium is mixed. Fracture risk with antidepressants is associated with duration of use, rather than dosage.

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

CASE Broken down
Ms. E, age 20, is a college student who has had major depressive disorder for several years and a genetic bone disease (osteogenesis imperfecta, mixed type III and IV). She pres­ents with depression, anxiety, and suicidal ideation. She reports recent worsening of her depressive symptoms, including anhedonia, excessive sleep, difficulty concentrating, and feeling overwhelmed, hopeless, and worth­less. She also describes frequent thoughts of suicide with the plan of putting herself in oncoming traffic, although she has no history of suicide attempts.

Previously, her primary care physician prescribed lorazepam, 0.5 mg, as needed for anxiety, and sertraline, 100 mg/d, for depres­sion and anxiety. She experienced only partial improvement in symptoms, however.

In addition to depressive symptoms, Ms. E describes manic symptoms lasting for as long as 3 to 5 days, including decreased need for sleep, increased energy, pressured speech, racing thoughts, distractibility, spending excessive money on cosmetics, and risking her safety—given her skeletal disorder— by participating in high-impact stage-combat classes. She denies auditory and visual hallucinations, homicidal ideation, and delusions.

The medical history is significant for osteogenesis imperfecta, which has caused 62 fractures and required 16 surgeries. Ms. E is a theater major who, despite her short stature and wheelchair use, reports enjoying her acting career and says she does not feel demoralized by her medical condition. She describes overcoming her physical disabili­ties with pride and confidence. However, her recent worsening mood symptoms have left her unable to concentrate and feeling over­whelmed with school.

Ms. E is voluntarily admitted to an inpatient psychiatric unit with a diagnosis of bipolar I disorder with rapid cycling, most recent epi­sode mixed. Because of her bone fragility, the treatment team considers what would be an appropriate course of drug treatment to con­trol bipolar symptoms while minimizing risk of bone loss.

Which medications are associated with decreased bone mineral density?
   a) citalopram
   b) haloperidol
   c) carbamazepine
   d) paliperidone
   e) all of the above

The authors’ observations
Osteogenesis imperfecta is a genetic condi­tion caused by mutations in genes impli­cated in collagen production. As a result, bones are brittle and prone to fracture. Different classes of psychotropics have been shown to increase risk of bone frac­tures through a variety of mechanisms. Clinicians often must choose appropri­ate pharmacotherapy for patients at high risk of fracture, including postmenopausal women, older patients, malnourished per­sons, and those with hormonal deficien­cies leading to osteoporosis.

To assist our clinical decision-making, we reviewed the literature to establish appropriate management of a patient with increased bone fragility and new-onset bipolar disorder. We considered all classes of medications used to treat bipolar disor­der, including antipsychotics, antidepres­sants, lithium, and anticonvulsants.

Antipsychotics
In population-based studies, prolactin-elevating antipsychotics have been associ­ated with decreased bone mineral density and increased risk of fracture.¹ Additional studies on geriatric and non-geriatric popu­lations have supported these findings.^2,3

The mechanism through which fracture risk is increased likely is related to antipsy­chotics’ effect on serum prolactin and corti­sol levels. Antipsychotics act as antagonists on D2 receptors in the hypothalamic tubero-infundibular pathway, therefore preventing inhibition of prolactin. Long-term elevation in serum prolactin can cause loss of bone mineral density through secondary hypo­gonadism and direct effects on target tis­sues. Additional modifying factors include smoking and estrogen use.

The degree to which antipsychotics increase fracture risk might be related to the degree of serum prolactin elevation.⁴ Antipsychotics previously have been grouped by the degree of prolactin eleva­tion, categorizing them as high, medium, and low or no potential to elevate serum prolactin.⁴ Based on this classification, typical antipsychotics, risperidone, and paliperidone have the highest potential to elevate prolactin. Accordingly, antipsychot­ics with the lowest fracture risk are those that have the lowest risk of serum prolactin elevation: ziprasidone, asenapine, quetiap­ine, and clozapine. Aripiprazole may lower prolactin in some patients. This is sup­ported by studies noting reduced bone min­eral density^5,6 and increased risk of fracture¹ with high-potential vs low- or no-potential antipsychotics. Because of these findings, it is crucial to consider the potential risk of prolactin elevation when treating patients at increased risk of fracture. Providers should consider low/no potential antipsy­chotic medications before considering those with medium or high potential (Table).

Antidepressants
In a meta-analysis, antidepressants were shown to increase fracture risk by 70% to 90%.² However, the relative risk varies by antidepressant class. Several studies have shown that selective serotonin reuptake inhibitors (SSRIs) are associated with a higher risk of fracture compared with tri­cyclic antidepressants (TCAs).⁷ In addition, antidepressants with a high affinity for the serotonin transporter, including citalopram, fluoxetine, fluvoxamine, paroxetine, sertra­line, and imipramine, have been associated with greater risk of osteoporotic fracture compared with those with low affinity.⁸

The mechanisms by which antidepres­sants increase fracture risk are complex, although the strongest evidence implicates a direct effect on bone metabolism via the 5-HTT receptor. This receptor, found on osteoblasts and osteoclasts, plays an impor­tant role in bone metabolism; it is through this receptor that SSRIs might inhibit osteo­blasts and promote osteoclast activity, thereby disrupting bone microarchitecture. Additional studies are needed to further describe the mechanism of the association among antidepressants, bone mineral den­sity, and fracture risk.

Fracture risk is associated with dura­tion of use rather than dosage. Population-based studies show a higher fracture risk for new users of TCAs compared with continuous users, and the risk of fracture with SSRIs seems to increase slightly over time.⁹ No association has been identified between fracture risk and antidepressant dosage. According to the literature, drugs with low affinity for the serotonin trans­porter, such as maprotiline and mirtazap­ine, likely are the safest antidepressants for patients at increased risk of fracture. Options also include other TCAs and any antidepressant with low affinity for the serotonin receptor.^7,8 

Lithium
Studies on lithium and bone mineral den­sity have shown mixed results. Older stud­ies found that lithium had a negative or no effect on bone mineral density or the para­thyroid hormone level.¹⁰ More recent inves­tigations, however, suggest that the drug has a protective effect on bone mineral den­sity, although this has not been replicated in all studies.

In a mouse model, lithium has been shown to enhance bone formation and improve bone mass, at least in part by activation of the Wnt signaling pathway through an inhibitory effect on glycogen synthase kinase-3β.¹¹ In humans, lithium-treated adults had lower serum alkaline phosphate, osteocalcin, and C-telopeptide levels compared with controls, suggesting a state of decreased bone remodeling and increased turnover.¹² There is a paucity of clinical data on the effect of lithium on frac­ture risk. Additional studies are necessary to elucidate lithium’s mechanism on bone mineral density and determine the magni­tude of the clinical effect.

Anticonvulsants
The association among anticonvulsants, decreased bone mineral density, and increased risk of fracture is well-established in the literature.¹³ However, causality is dif­ficult to determine, because many studies were of patients with a seizure disorder, who often have additional risk factors for fracture, including seizure-related trauma, drowsiness, and slowed reflexes.

Mechanisms through which anticon­vulsants increase fracture risk include increased bone resorption, secondary hypo­parathyroidism, and pseudohypoparathy­roidism. Markers of bone resorption were elevated in patients receiving an antiepi­leptic.¹⁴ This effect might be enhanced by co-administration of cytochrome P450 (CYP450) enzyme-inducing anticon­vulsants and CYP450 enzyme-inhibiting medications, such as valproate. Long-term treatment with valproate may produce reduction of bone mass and increased risk of fractures; however, other studies dis­agree with this finding.¹⁵

In addition to CYP450-inducing effects, phenytoin, carbamezapine, and phenobar­bital can increase catabolism of vitamin D, which is associated with osteomalacia.¹⁴ This results in decreased intestinal absorp­tion of calcium, hypocalcemia, and sec­ondary hyperparathyroidism, which also increases fracture risk. Anticonvulsants also might increase resistance to pseudo­hypoparathyroidism and inhibit calcitonin secretion.

Lamotrigine has not been shown to interfere with bone accrual¹⁶ and may be a safer mood stabilizer for patients at high risk of fracture. For patients at increased risk of fracture, it is important to select an anticonvulsant wisely to minimize frac­ture risk.

How would you treat Ms. E during her hospitalization for bipolar disorder?
   a) carbamazepine
   b) lithium
   c) risperidone
   d) mirtazapine

TREATMENT Minimizing polypharmacy
Because many pharmacotherapeutic options for managing bipolar disorder can increase the risk of fracture, clinicians must be aware of the relative risk of each class of medication and each individual drug. We initiated lithium, 300 mg, 3 times a day, to stabilize Ms. E’s mood. Although clinical data are inconclusive regarding lithium’s effect on fracture risk, we felt that the benefit of acute mood stabiliza­tion outweighed the risk of decreased bone mineral index.

We selected aripiprazole, 10 mg/d, as an adjunctive treatment because of its minimal effect on serum prolactin levels.⁴ We con­sidered prescribing an antidepressant but decided against it because we were concerned about manic switching.

Polypharmacy is another important con­sideration for Ms. E. Several studies have identified polypharmacy, particularly with antipsychotics, as an independent risk factor for fracture.³ Therefore, we sought to minimize the number of medications Ms. E receives. Although lithium monotherapy is an option, we thought that her mood symptoms were severe enough that the risk of inadequately treating her bipolar symptoms outweighed the additional risk of fracture from dual ther­apy with lithium and aripiprazole. Untreated or inadequately treated depression is associ­ated with a higher fracture risk. Therefore, we avoided prescribing >2 medications to mitigate any excessive risk of fracture from polypharmacy.

Bottom Line
Different classes of medications—antipsychotics, anticonvulsants, antidepressants, and lithium—used for treating bipolar disorder have been shown to increase risk of bone fracture through a variety of mechanisms. Anticonvulsants and prolactin-elevating antipsychotics are associated with increased fracture risk; evidence on lithium is mixed. Fracture risk with antidepressants is associated with duration of use, rather than dosage.

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

CASE Broken down
Ms. E, age 20, is a college student who has had major depressive disorder for several years and a genetic bone disease (osteogenesis imperfecta, mixed type III and IV). She pres­ents with depression, anxiety, and suicidal ideation. She reports recent worsening of her depressive symptoms, including anhedonia, excessive sleep, difficulty concentrating, and feeling overwhelmed, hopeless, and worth­less. She also describes frequent thoughts of suicide with the plan of putting herself in oncoming traffic, although she has no history of suicide attempts.

Previously, her primary care physician prescribed lorazepam, 0.5 mg, as needed for anxiety, and sertraline, 100 mg/d, for depres­sion and anxiety. She experienced only partial improvement in symptoms, however.

In addition to depressive symptoms, Ms. E describes manic symptoms lasting for as long as 3 to 5 days, including decreased need for sleep, increased energy, pressured speech, racing thoughts, distractibility, spending excessive money on cosmetics, and risking her safety—given her skeletal disorder— by participating in high-impact stage-combat classes. She denies auditory and visual hallucinations, homicidal ideation, and delusions.

The medical history is significant for osteogenesis imperfecta, which has caused 62 fractures and required 16 surgeries. Ms. E is a theater major who, despite her short stature and wheelchair use, reports enjoying her acting career and says she does not feel demoralized by her medical condition. She describes overcoming her physical disabili­ties with pride and confidence. However, her recent worsening mood symptoms have left her unable to concentrate and feeling over­whelmed with school.

Ms. E is voluntarily admitted to an inpatient psychiatric unit with a diagnosis of bipolar I disorder with rapid cycling, most recent epi­sode mixed. Because of her bone fragility, the treatment team considers what would be an appropriate course of drug treatment to con­trol bipolar symptoms while minimizing risk of bone loss.

Which medications are associated with decreased bone mineral density?
   a) citalopram
   b) haloperidol
   c) carbamazepine
   d) paliperidone
   e) all of the above

The authors’ observations
Osteogenesis imperfecta is a genetic condi­tion caused by mutations in genes impli­cated in collagen production. As a result, bones are brittle and prone to fracture. Different classes of psychotropics have been shown to increase risk of bone frac­tures through a variety of mechanisms. Clinicians often must choose appropri­ate pharmacotherapy for patients at high risk of fracture, including postmenopausal women, older patients, malnourished per­sons, and those with hormonal deficien­cies leading to osteoporosis.

To assist our clinical decision-making, we reviewed the literature to establish appropriate management of a patient with increased bone fragility and new-onset bipolar disorder. We considered all classes of medications used to treat bipolar disor­der, including antipsychotics, antidepres­sants, lithium, and anticonvulsants.

Antipsychotics
In population-based studies, prolactin-elevating antipsychotics have been associ­ated with decreased bone mineral density and increased risk of fracture.¹ Additional studies on geriatric and non-geriatric popu­lations have supported these findings.^2,3

The mechanism through which fracture risk is increased likely is related to antipsy­chotics’ effect on serum prolactin and corti­sol levels. Antipsychotics act as antagonists on D2 receptors in the hypothalamic tubero-infundibular pathway, therefore preventing inhibition of prolactin. Long-term elevation in serum prolactin can cause loss of bone mineral density through secondary hypo­gonadism and direct effects on target tis­sues. Additional modifying factors include smoking and estrogen use.

The degree to which antipsychotics increase fracture risk might be related to the degree of serum prolactin elevation.⁴ Antipsychotics previously have been grouped by the degree of prolactin eleva­tion, categorizing them as high, medium, and low or no potential to elevate serum prolactin.⁴ Based on this classification, typical antipsychotics, risperidone, and paliperidone have the highest potential to elevate prolactin. Accordingly, antipsychot­ics with the lowest fracture risk are those that have the lowest risk of serum prolactin elevation: ziprasidone, asenapine, quetiap­ine, and clozapine. Aripiprazole may lower prolactin in some patients. This is sup­ported by studies noting reduced bone min­eral density^5,6 and increased risk of fracture¹ with high-potential vs low- or no-potential antipsychotics. Because of these findings, it is crucial to consider the potential risk of prolactin elevation when treating patients at increased risk of fracture. Providers should consider low/no potential antipsy­chotic medications before considering those with medium or high potential (Table).

Antidepressants
In a meta-analysis, antidepressants were shown to increase fracture risk by 70% to 90%.² However, the relative risk varies by antidepressant class. Several studies have shown that selective serotonin reuptake inhibitors (SSRIs) are associated with a higher risk of fracture compared with tri­cyclic antidepressants (TCAs).⁷ In addition, antidepressants with a high affinity for the serotonin transporter, including citalopram, fluoxetine, fluvoxamine, paroxetine, sertra­line, and imipramine, have been associated with greater risk of osteoporotic fracture compared with those with low affinity.⁸

The mechanisms by which antidepres­sants increase fracture risk are complex, although the strongest evidence implicates a direct effect on bone metabolism via the 5-HTT receptor. This receptor, found on osteoblasts and osteoclasts, plays an impor­tant role in bone metabolism; it is through this receptor that SSRIs might inhibit osteo­blasts and promote osteoclast activity, thereby disrupting bone microarchitecture. Additional studies are needed to further describe the mechanism of the association among antidepressants, bone mineral den­sity, and fracture risk.

Fracture risk is associated with dura­tion of use rather than dosage. Population-based studies show a higher fracture risk for new users of TCAs compared with continuous users, and the risk of fracture with SSRIs seems to increase slightly over time.⁹ No association has been identified between fracture risk and antidepressant dosage. According to the literature, drugs with low affinity for the serotonin trans­porter, such as maprotiline and mirtazap­ine, likely are the safest antidepressants for patients at increased risk of fracture. Options also include other TCAs and any antidepressant with low affinity for the serotonin receptor.^7,8 

Lithium
Studies on lithium and bone mineral den­sity have shown mixed results. Older stud­ies found that lithium had a negative or no effect on bone mineral density or the para­thyroid hormone level.¹⁰ More recent inves­tigations, however, suggest that the drug has a protective effect on bone mineral den­sity, although this has not been replicated in all studies.

In a mouse model, lithium has been shown to enhance bone formation and improve bone mass, at least in part by activation of the Wnt signaling pathway through an inhibitory effect on glycogen synthase kinase-3β.¹¹ In humans, lithium-treated adults had lower serum alkaline phosphate, osteocalcin, and C-telopeptide levels compared with controls, suggesting a state of decreased bone remodeling and increased turnover.¹² There is a paucity of clinical data on the effect of lithium on frac­ture risk. Additional studies are necessary to elucidate lithium’s mechanism on bone mineral density and determine the magni­tude of the clinical effect.

Anticonvulsants
The association among anticonvulsants, decreased bone mineral density, and increased risk of fracture is well-established in the literature.¹³ However, causality is dif­ficult to determine, because many studies were of patients with a seizure disorder, who often have additional risk factors for fracture, including seizure-related trauma, drowsiness, and slowed reflexes.

Mechanisms through which anticon­vulsants increase fracture risk include increased bone resorption, secondary hypo­parathyroidism, and pseudohypoparathy­roidism. Markers of bone resorption were elevated in patients receiving an antiepi­leptic.¹⁴ This effect might be enhanced by co-administration of cytochrome P450 (CYP450) enzyme-inducing anticon­vulsants and CYP450 enzyme-inhibiting medications, such as valproate. Long-term treatment with valproate may produce reduction of bone mass and increased risk of fractures; however, other studies dis­agree with this finding.¹⁵

In addition to CYP450-inducing effects, phenytoin, carbamezapine, and phenobar­bital can increase catabolism of vitamin D, which is associated with osteomalacia.¹⁴ This results in decreased intestinal absorp­tion of calcium, hypocalcemia, and sec­ondary hyperparathyroidism, which also increases fracture risk. Anticonvulsants also might increase resistance to pseudo­hypoparathyroidism and inhibit calcitonin secretion.

Lamotrigine has not been shown to interfere with bone accrual¹⁶ and may be a safer mood stabilizer for patients at high risk of fracture. For patients at increased risk of fracture, it is important to select an anticonvulsant wisely to minimize frac­ture risk.

How would you treat Ms. E during her hospitalization for bipolar disorder?
   a) carbamazepine
   b) lithium
   c) risperidone
   d) mirtazapine

TREATMENT Minimizing polypharmacy
Because many pharmacotherapeutic options for managing bipolar disorder can increase the risk of fracture, clinicians must be aware of the relative risk of each class of medication and each individual drug. We initiated lithium, 300 mg, 3 times a day, to stabilize Ms. E’s mood. Although clinical data are inconclusive regarding lithium’s effect on fracture risk, we felt that the benefit of acute mood stabiliza­tion outweighed the risk of decreased bone mineral index.

We selected aripiprazole, 10 mg/d, as an adjunctive treatment because of its minimal effect on serum prolactin levels.⁴ We con­sidered prescribing an antidepressant but decided against it because we were concerned about manic switching.

Polypharmacy is another important con­sideration for Ms. E. Several studies have identified polypharmacy, particularly with antipsychotics, as an independent risk factor for fracture.³ Therefore, we sought to minimize the number of medications Ms. E receives. Although lithium monotherapy is an option, we thought that her mood symptoms were severe enough that the risk of inadequately treating her bipolar symptoms outweighed the additional risk of fracture from dual ther­apy with lithium and aripiprazole. Untreated or inadequately treated depression is associ­ated with a higher fracture risk. Therefore, we avoided prescribing >2 medications to mitigate any excessive risk of fracture from polypharmacy.

Bottom Line
Different classes of medications—antipsychotics, anticonvulsants, antidepressants, and lithium—used for treating bipolar disorder have been shown to increase risk of bone fracture through a variety of mechanisms. Anticonvulsants and prolactin-elevating antipsychotics are associated with increased fracture risk; evidence on lithium is mixed. Fracture risk with antidepressants is associated with duration of use, rather than dosage.

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