Reviews

Credit cards have made it easier for psychiatrists who work in outpatient settings to collect payment for their services. Accepting credit cards saves time in sessions for clinical matters, leads to higher rates of collecting payments for patients who do not show up for appointments, and avoids having to manage bounced checks and collection agencies.¹ No federal or state laws prohibit businesses from storing consumers’ credit card information. However, psychiatric practices are legally obligated to have safeguards in place to protect sensitive information and limit liability exposures.² There are several steps to take when storing your patients’ credit card information.

Establish a payment policy. Create a policy that outlines your practice’s credit card procedures, including when credit cards will be charged and under what circumstances, how patients will be notified, and how credit card information will be stored.² Give your patients a copy of this policy and review it with them at their first appointment and any time you change this policy.² Get consent from your patients before using and storing their credit card information.²

Use secure methods to store this information. Most medical practices photocopy/write down their patients’ credit card information and store it in the patient’s electronic/paper medical record, or they use an online service to store it electronically.² Online services usually provide a higher level of protection than the patient’s medical record.² Ensure that electronic data that includes credit card numbers is robustly encrypted, or that paper records are locked in a secure place, such as in a safe or file drawer that requires a key/combination lock.³ Payment Card Industry (PCI) regulations prohibit storing a credit card’s security code (a three- or four-digit number on the front or back of the card).³ This code is used to allow merchants to verify whether a customer authorizing a transaction over the phone or via the internet physically possesses the card.³ PCI regulations also prohibit storing data contained in the card’s magnetic strip.³ This data contains information about the account that is not displayed on the card, assists with authorizing transactions, and ensures that credit cards cannot be easily counterfeited.³

Understand all federal and state laws and regulations. If your practice collects patient billing information, you are considered a “merchant” and are subject to federal and state laws and regulations that protect consumer credit card information.² These laws and regulations include (but are not limited to)²:

• Health Insurance Portability and Accountability Act (HIPAA) and similar state privacy laws
• Federal Trade Commission Act (FTCA) and similar state business laws
• Payment Card Industry Data Security Standard (PCI DSS), which was not devised by federal or state government.

HIPAA and state privacy laws require psychiatrists to implement “reasonable” security measures to protect payment information, regardless of how that information is stored.^2,4 Because HIPAA does not define “reasonable,” psychiatrists have latitude in determining which security measures to implement.^2,4 Locking the information in a file cabinet and locking the room where the file cabinet is kept (for paper storage) or using HIPAA-compliant encrypted storage programs (for electronic storage) are examples of “reasonable” security measures.²

FTCA requires businesses to use “appropriate” and “reasonable” security measures to protect credit card information.^2,5 Because FTCA does not specify these terms,^2,5 psychiatrists have leeway in determining which security measures to implement. Federal law requires all businesses to delete a card’s expiration date and shorten the account information to include no more than the last 5 digits of the card number that is printed on all sales receipts.⁶ FTCA also requires businesses to get prior authorization from individuals before charging their credit card.² For example, if a patient previously used a credit card to pay for a session, the psychiatrist cannot later use the credit card to charge for a missed appointment without notifying the patient and getting their authorization.²

PCI DSS applies to entities that store, process, and/or transmit cardholder data.⁷ Any business that accepts credit card payments must comply with PCI DSS, which includes 12 requirements.⁷ Examples of these requirements include using firewalls to protect cardholder data and restricting access to cardholder data to a “need-to-know” basis. Businesses that do not comply with PCI DSS can be subjected to fines and/or have their contracts terminated by the credit card companies.² Fines can range from $5,000 to $100,000 per month for data breaches where you are found negligible.¹

Credit cards have made it easier for psychiatrists who work in outpatient settings to collect payment for their services. Accepting credit cards saves time in sessions for clinical matters, leads to higher rates of collecting payments for patients who do not show up for appointments, and avoids having to manage bounced checks and collection agencies.¹ No federal or state laws prohibit businesses from storing consumers’ credit card information. However, psychiatric practices are legally obligated to have safeguards in place to protect sensitive information and limit liability exposures.² There are several steps to take when storing your patients’ credit card information.

Establish a payment policy. Create a policy that outlines your practice’s credit card procedures, including when credit cards will be charged and under what circumstances, how patients will be notified, and how credit card information will be stored.² Give your patients a copy of this policy and review it with them at their first appointment and any time you change this policy.² Get consent from your patients before using and storing their credit card information.²

Use secure methods to store this information. Most medical practices photocopy/write down their patients’ credit card information and store it in the patient’s electronic/paper medical record, or they use an online service to store it electronically.² Online services usually provide a higher level of protection than the patient’s medical record.² Ensure that electronic data that includes credit card numbers is robustly encrypted, or that paper records are locked in a secure place, such as in a safe or file drawer that requires a key/combination lock.³ Payment Card Industry (PCI) regulations prohibit storing a credit card’s security code (a three- or four-digit number on the front or back of the card).³ This code is used to allow merchants to verify whether a customer authorizing a transaction over the phone or via the internet physically possesses the card.³ PCI regulations also prohibit storing data contained in the card’s magnetic strip.³ This data contains information about the account that is not displayed on the card, assists with authorizing transactions, and ensures that credit cards cannot be easily counterfeited.³

Understand all federal and state laws and regulations. If your practice collects patient billing information, you are considered a “merchant” and are subject to federal and state laws and regulations that protect consumer credit card information.² These laws and regulations include (but are not limited to)²:

• Health Insurance Portability and Accountability Act (HIPAA) and similar state privacy laws
• Federal Trade Commission Act (FTCA) and similar state business laws
• Payment Card Industry Data Security Standard (PCI DSS), which was not devised by federal or state government.

HIPAA and state privacy laws require psychiatrists to implement “reasonable” security measures to protect payment information, regardless of how that information is stored.^2,4 Because HIPAA does not define “reasonable,” psychiatrists have latitude in determining which security measures to implement.^2,4 Locking the information in a file cabinet and locking the room where the file cabinet is kept (for paper storage) or using HIPAA-compliant encrypted storage programs (for electronic storage) are examples of “reasonable” security measures.²

FTCA requires businesses to use “appropriate” and “reasonable” security measures to protect credit card information.^2,5 Because FTCA does not specify these terms,^2,5 psychiatrists have leeway in determining which security measures to implement. Federal law requires all businesses to delete a card’s expiration date and shorten the account information to include no more than the last 5 digits of the card number that is printed on all sales receipts.⁶ FTCA also requires businesses to get prior authorization from individuals before charging their credit card.² For example, if a patient previously used a credit card to pay for a session, the psychiatrist cannot later use the credit card to charge for a missed appointment without notifying the patient and getting their authorization.²

PCI DSS applies to entities that store, process, and/or transmit cardholder data.⁷ Any business that accepts credit card payments must comply with PCI DSS, which includes 12 requirements.⁷ Examples of these requirements include using firewalls to protect cardholder data and restricting access to cardholder data to a “need-to-know” basis. Businesses that do not comply with PCI DSS can be subjected to fines and/or have their contracts terminated by the credit card companies.² Fines can range from $5,000 to $100,000 per month for data breaches where you are found negligible.¹

Credit cards have made it easier for psychiatrists who work in outpatient settings to collect payment for their services. Accepting credit cards saves time in sessions for clinical matters, leads to higher rates of collecting payments for patients who do not show up for appointments, and avoids having to manage bounced checks and collection agencies.¹ No federal or state laws prohibit businesses from storing consumers’ credit card information. However, psychiatric practices are legally obligated to have safeguards in place to protect sensitive information and limit liability exposures.² There are several steps to take when storing your patients’ credit card information.

Establish a payment policy. Create a policy that outlines your practice’s credit card procedures, including when credit cards will be charged and under what circumstances, how patients will be notified, and how credit card information will be stored.² Give your patients a copy of this policy and review it with them at their first appointment and any time you change this policy.² Get consent from your patients before using and storing their credit card information.²

Use secure methods to store this information. Most medical practices photocopy/write down their patients’ credit card information and store it in the patient’s electronic/paper medical record, or they use an online service to store it electronically.² Online services usually provide a higher level of protection than the patient’s medical record.² Ensure that electronic data that includes credit card numbers is robustly encrypted, or that paper records are locked in a secure place, such as in a safe or file drawer that requires a key/combination lock.³ Payment Card Industry (PCI) regulations prohibit storing a credit card’s security code (a three- or four-digit number on the front or back of the card).³ This code is used to allow merchants to verify whether a customer authorizing a transaction over the phone or via the internet physically possesses the card.³ PCI regulations also prohibit storing data contained in the card’s magnetic strip.³ This data contains information about the account that is not displayed on the card, assists with authorizing transactions, and ensures that credit cards cannot be easily counterfeited.³

Understand all federal and state laws and regulations. If your practice collects patient billing information, you are considered a “merchant” and are subject to federal and state laws and regulations that protect consumer credit card information.² These laws and regulations include (but are not limited to)²:

• Health Insurance Portability and Accountability Act (HIPAA) and similar state privacy laws
• Federal Trade Commission Act (FTCA) and similar state business laws
• Payment Card Industry Data Security Standard (PCI DSS), which was not devised by federal or state government.

HIPAA and state privacy laws require psychiatrists to implement “reasonable” security measures to protect payment information, regardless of how that information is stored.^2,4 Because HIPAA does not define “reasonable,” psychiatrists have latitude in determining which security measures to implement.^2,4 Locking the information in a file cabinet and locking the room where the file cabinet is kept (for paper storage) or using HIPAA-compliant encrypted storage programs (for electronic storage) are examples of “reasonable” security measures.²

FTCA requires businesses to use “appropriate” and “reasonable” security measures to protect credit card information.^2,5 Because FTCA does not specify these terms,^2,5 psychiatrists have leeway in determining which security measures to implement. Federal law requires all businesses to delete a card’s expiration date and shorten the account information to include no more than the last 5 digits of the card number that is printed on all sales receipts.⁶ FTCA also requires businesses to get prior authorization from individuals before charging their credit card.² For example, if a patient previously used a credit card to pay for a session, the psychiatrist cannot later use the credit card to charge for a missed appointment without notifying the patient and getting their authorization.²

PCI DSS applies to entities that store, process, and/or transmit cardholder data.⁷ Any business that accepts credit card payments must comply with PCI DSS, which includes 12 requirements.⁷ Examples of these requirements include using firewalls to protect cardholder data and restricting access to cardholder data to a “need-to-know” basis. Businesses that do not comply with PCI DSS can be subjected to fines and/or have their contracts terminated by the credit card companies.² Fines can range from $5,000 to $100,000 per month for data breaches where you are found negligible.¹

Cigarette smoking is common among patients with schizophrenia, mood disorders, anxiety disorders,^1-3 substance use disorders (SUDs),⁴ and other psychiatric disorders. Research suggests that compared with the general population, patients with SUDs consume more nicotine products and are more vulnerable to the effects of smoking.⁵ Despite the availability of effective treatments, many mental health professionals are reluctant to identify and treat tobacco use disorder,^6-8 or they prioritize other disorders over tobacco use. Early detection and treatment of tobacco use disorder can improve patients’ health and reduce the incidence of acute and chronic illness.

Cotinine is a biomarker that can be used to detect tobacco use. It can be measured in routine clinical practice by collecting urinary, serum, or salivary specimens, and used to monitor psychiatric patients’ tobacco use. Monitoring cotinine levels is similar to using other biomarkers to assess medication adherence or identify illicit substance use. A growing body of evidence supports the utility of cotinine screening as a part of a comprehensive substance use disorder treatment plan,^5,9,10 especially for:

• patients who have comorbid conditions that can be exacerbated by tobacco use, such as chronic obstructive pulmonary disease
• patients who are pregnant^11,12
• patients who are less reliable in self-report or who require objective testing for validation.

Routine clinical screening of tobacco use is recommended for all patients and early detection may facilitate earlier treatment. Several FDA-approved medications are available for smoking cessation¹³; however, discussion of treatment options is beyond the scope of this review. In this article, we describe how cotinine is measured and analyzed, 3 case vignettes that illustrate its potential clinical utility, and limitations to its use as a biomarker of tobacco use.

Methods of measuring cotinine

Cigarette smoking is associated with the absorption of nicotine, which is mainly metabolized by cytochrome P450 (CYP) 2A6 to 6 primary metabolites: cotinine, hydroxycotinine, norcotinine, nornicotine, cotinine oxide, and nicotine oxide.^14,15 Cotinine is the biomarker of choice for detecting use of tobacco/nicotine products due to its stability (it is not influenced by dietary or environmental factors), extended half-life (16 to 19 hours, compared with 2 hours for nicotine), and stable concentration throughout the day. Samples from saliva, urine, or blood can be analyzed through radioimmunoassay, enzyme-linked immunosorbent assay (ELISA), and gas/liquid chromatography.¹⁶ The specificity of cotinine for tobacco use is excellent, except for persons who are taking medications that contain nicotine.¹⁷

An advantage of cotinine over other biomarkers for smoking (such as carbon monoxide in expired air) is that the optimal cut-off points for cotinine are relatively uninfluenced by the prevalence of smoking in the population. The optimal cut-off levels used to detect current tobacco use may vary based on the sample or test used (saliva, urine, or plasma) and certain patient-specific factors (Box 1^11,16,18-21). However, for plasma or saliva cotinine, 16 ng/mL is the generally accepted cut-off level for detecting current tobacco use. A urinary cotinine cut-off level of 50 ng/mL is likely appropriate for most circumstances.¹⁷ Users of electronic nicotine delivery systems (electronic cigarettes) have been found to have cotinine levels similar to those of cigarette smokers.²²

Box 1

Assessment of cotinine levels in saliva may be considered for outpatient monitoring due to its noninvasive nature, tolerability, and the ability to collect multiple samples over a limited period.²³ Saliva cotinine levels correlate closely with blood concentrations. Urine cotinine levels offer some advantage because concentrations are 6 times higher in urine than in blood or saliva. For this reason, urine cotinine is the most widely used biomarker in individuals who use tobacco due to its high sensitivity, specificity, reliability, and noninvasive collection.²³ By using a lower urinary cut-off of ≥2.47 ng/mL, ELISA kits detect the highest sensitivity and specificity, which is useful for monitoring daily tobacco use.²⁴ This cut-off value was associated with 100% sensitivity and specificity, and these numbers declined with increases in the cut-off threshold.²³

Impact of cigarette smoking

The following 3 clinical vignettes illustrate the impact of tobacco use disorder on patients, and how cotinine might help with their treatment.

Continue to: Vignette 1

Mr. D, age 44, has a history of schizophrenia and has smoked 1 pack of cigarettes per day for the last 15 years. He was recently discharged from an inpatient psychiatric facility after his symptoms were stabilized. During his hospitalization, Mr. D used a nicotine-replacement product to comply with the hospital’s smoke-free policy. Unfortunately, since discharge, Mr. D reports worsening auditory hallucinations despite adherence with his antipsychotic medication, clozapine, 600 mg at bedtime. Collateral information gathered from Mr. D’s mother confirms that he has been adherent with the discharge medication regimen; however, Mr. D has resumed smoking 1 pack of cigarettes daily. The treatment team suspects that his worsening psychosis is related to the decrease of blood clozapine level due to CYP induction by cigarette smoke.

Cotinine and smoking-related drug interactions

Vignette 1 illustrates the significant impact tobacco smoke can have on the effectiveness of a psychotropic medication. This is caused by polycyclic aromatic hydrocarbons induction of hepatic CYP1A2 isoenzymes. Clinicians should routinely screen patients for smoking status due to the potential for drug interactions. Common major CYP1A2 substrates include clozapine, olanzapine, duloxetine, and mirtazapine. Common minor CYP1A2 substrates include asenapine, chlorpromazine, haloperidol, perphenazine, ziprasidone, and zolpidem.²⁵ This case also highlights the potential advantage of collecting collateral information, and the utility of objective laboratory testing (such as cotinine monitoring) for optimal treatment because patient self-report can be unreliable and incomplete.^12,26

Vignette 2

Mr. B, age 34, has a history of cocaine use disorder and tobacco use disorder. He is referred to a treatment program and participates in a contingency management program for his substance use disorders. Biomarkers, including salivary cotinine, are used to assess Mr. B’s exposure to tobacco use. Mr. B and other participants in his program are eligible for prize draws if they are found to have samples that are negative for tobacco and other substances. There are other incentives in place for patients who show a reduced cotinine concentration.

Cotinine monitoring and contingency management

Clinicians can incorporate cotinine monitoring into existing SUD treatment. This is similar to the utilization of other biomarkers that are commonly used to identify recent illicit substance use or monitor adherence to treatment medications. For example, benzoylecgonine, a metabolite of cocaine, is frequently used to monitor abstinence from cocaine.²⁷ In addition, buprenorphine metabolites are used to monitor medication adherence and the possibility of diversion among patients being treated for opioid use disorder.²⁸ Smoking cessation may be associated with a decreased risk of relapse to illicit substances.²⁹

Treatments based on contingency management principles involve giving patients tangible rewards to reinforce desired (positive) behaviors. Smoking cessation can be confirmed by monitoring cotinine levels. Gayman et al⁹ found twice-weekly salivary testing was compatible with monitoring and promoting abstinence in a prize-based contingency management smoking cessation program. Most prior studies used urine cotinine measures to verify abstinence. Although highly reliable, urine samples require close monitoring to ensure sample validity, which can be a burden on staff and unpleasant for patients.⁹ It is also important to note that the rate of elimination of cotinine from saliva and urine are comparable. The half-life of cotinine is approximately 18 hours, and therefore the specificity of salivary test strips may be impacted during the first 4 to 5 days of abstinence. In the first few days of smoking cessation, a more intensive approach, such as quantifying urine cotinine levels and monitoring decline, may be appropriate.²³

Continue to: Vignette 3

Vignette 3

Ms. C, age 34 and pregnant, is admitted to an outpatient treatment program for alcohol use disorder. She also has generalized anxiety disorder and tobacco use disorder. In addition to attending group therapy sessions and self-reporting any recent alcohol consumption, Ms. C also undergoes alcohol breathalyzer tests and urine studies of alcohol metabolites to monitor abstinence from alcohol. She says that the regular laboratory screening for alcohol use gives her a sense of accountability and tangible evidence of change that positively impacts her treatment. When the treating psychiatrist recommends that Ms. C also consider addressing her tobacco use disorder, she asks if there is some way to include laboratory testing to monitor her smoking cessation.

Cotinine as a predictor of smoking status

Smoking abstinence rates during pregnancy are lower than that for other substances, and pregnant women may not be aware of the impact of smoking on fetal development.³⁰ Cotinine can be used to verify self-report of smoking status and severity.^10,31,32

Salivary cotinine tests are commercially available, relatively economical, and convenient to use when frequent monitoring is required.³² In general, based on established cut-off values that are unique to the specimen collected, the overall high specificity and sensitivity of salivary testing allows clinicians to predict smoker vs nonsmoker status with confidence. For example, a 2008 study reported a salivary cotinine cut-off level of 12 ng/mL for smokers.²¹ The sensitivity and specificity of this cut-off value for distinguishing cigarette smokers from never smokers were 96.7% and 96.9%, respectively.²¹

Additionally, some studies suggest that cotinine levels may be predictive of treatment outcomes and retention in SUD treatment programs.^33,34 One study of smoking cessation using nicotine replacement products found that compared with patients with lower baseline cotinine levels prior to treatment, patients with higher baseline cotinine plasma levels had lower smoking cessation success rates.³⁴

A few caveats

There are several limitations to quantitative measures of cotinine (Box 2^21,23). These include (but are not limited to) potential errors related to sample collection, storage, shipping, and analysis.²³ Compared with other methods, point-of-care cotinine measurement in saliva is noninvasive, simple, and requires less training to properly use.²³

Box 2

Continue to: Confirmation of smoking cessation...

Confirmation of smoking cessation can be monitored reliably within the clinical setting using cotinine monitoring. However, this is not a routine test, and there are no guidelines or consensus on how or when it should be used. The clinical feasibility of cotinine monitoring for psychiatric patients will depend on the cost of testing, methods used, amount of reimbursement for performing the tests, and how clinicians value such testing.³⁵

Bottom Line

Cotinine is a biomarker that can be used to detect tobacco use. Cotinine measurement can be used to monitor tobacco use and smoking cessation in psychiatric patients. Early detection and treatment of tobacco use disorder can improve patients’ health and reduce the incidence of acute and chronic illnesses. However, cotinine measurement is not a routine test, and there are no guidelines on how or when this test should be used.

Related Resources

• Peckham E, Brabyn S, Cook L, et al. Smoking cessation in severe mental ill health: what works? An updated systematic review and meta-analysis. BMC Psychiatry. 2017;17(1):252.
• Tidey JW, Miller ME. Smoking cessation and reduction in people with chronic mental illness. BMJ. 2015;351:h4065. doi: 10.1136/bmj.h4065

Drug Brand Names

Asenapine • Saphris
Buprenorphine • Sublocade
Clozapine • Clozaril
Duloxetine • Cymbalta
Haloperidol • Haldol
Mirtazapine • Remeron
Olanzapine • Zyprexa
Ziprasidone • Geodon
Zolpidem • Ambien

Cigarette smoking is common among patients with schizophrenia, mood disorders, anxiety disorders,^1-3 substance use disorders (SUDs),⁴ and other psychiatric disorders. Research suggests that compared with the general population, patients with SUDs consume more nicotine products and are more vulnerable to the effects of smoking.⁵ Despite the availability of effective treatments, many mental health professionals are reluctant to identify and treat tobacco use disorder,^6-8 or they prioritize other disorders over tobacco use. Early detection and treatment of tobacco use disorder can improve patients’ health and reduce the incidence of acute and chronic illness.

Cotinine is a biomarker that can be used to detect tobacco use. It can be measured in routine clinical practice by collecting urinary, serum, or salivary specimens, and used to monitor psychiatric patients’ tobacco use. Monitoring cotinine levels is similar to using other biomarkers to assess medication adherence or identify illicit substance use. A growing body of evidence supports the utility of cotinine screening as a part of a comprehensive substance use disorder treatment plan,^5,9,10 especially for:

• patients who have comorbid conditions that can be exacerbated by tobacco use, such as chronic obstructive pulmonary disease
• patients who are pregnant^11,12
• patients who are less reliable in self-report or who require objective testing for validation.

Routine clinical screening of tobacco use is recommended for all patients and early detection may facilitate earlier treatment. Several FDA-approved medications are available for smoking cessation¹³; however, discussion of treatment options is beyond the scope of this review. In this article, we describe how cotinine is measured and analyzed, 3 case vignettes that illustrate its potential clinical utility, and limitations to its use as a biomarker of tobacco use.

Methods of measuring cotinine

Cigarette smoking is associated with the absorption of nicotine, which is mainly metabolized by cytochrome P450 (CYP) 2A6 to 6 primary metabolites: cotinine, hydroxycotinine, norcotinine, nornicotine, cotinine oxide, and nicotine oxide.^14,15 Cotinine is the biomarker of choice for detecting use of tobacco/nicotine products due to its stability (it is not influenced by dietary or environmental factors), extended half-life (16 to 19 hours, compared with 2 hours for nicotine), and stable concentration throughout the day. Samples from saliva, urine, or blood can be analyzed through radioimmunoassay, enzyme-linked immunosorbent assay (ELISA), and gas/liquid chromatography.¹⁶ The specificity of cotinine for tobacco use is excellent, except for persons who are taking medications that contain nicotine.¹⁷

An advantage of cotinine over other biomarkers for smoking (such as carbon monoxide in expired air) is that the optimal cut-off points for cotinine are relatively uninfluenced by the prevalence of smoking in the population. The optimal cut-off levels used to detect current tobacco use may vary based on the sample or test used (saliva, urine, or plasma) and certain patient-specific factors (Box 1^11,16,18-21). However, for plasma or saliva cotinine, 16 ng/mL is the generally accepted cut-off level for detecting current tobacco use. A urinary cotinine cut-off level of 50 ng/mL is likely appropriate for most circumstances.¹⁷ Users of electronic nicotine delivery systems (electronic cigarettes) have been found to have cotinine levels similar to those of cigarette smokers.²²

Box 1

Assessment of cotinine levels in saliva may be considered for outpatient monitoring due to its noninvasive nature, tolerability, and the ability to collect multiple samples over a limited period.²³ Saliva cotinine levels correlate closely with blood concentrations. Urine cotinine levels offer some advantage because concentrations are 6 times higher in urine than in blood or saliva. For this reason, urine cotinine is the most widely used biomarker in individuals who use tobacco due to its high sensitivity, specificity, reliability, and noninvasive collection.²³ By using a lower urinary cut-off of ≥2.47 ng/mL, ELISA kits detect the highest sensitivity and specificity, which is useful for monitoring daily tobacco use.²⁴ This cut-off value was associated with 100% sensitivity and specificity, and these numbers declined with increases in the cut-off threshold.²³

Impact of cigarette smoking

The following 3 clinical vignettes illustrate the impact of tobacco use disorder on patients, and how cotinine might help with their treatment.

Continue to: Vignette 1

Mr. D, age 44, has a history of schizophrenia and has smoked 1 pack of cigarettes per day for the last 15 years. He was recently discharged from an inpatient psychiatric facility after his symptoms were stabilized. During his hospitalization, Mr. D used a nicotine-replacement product to comply with the hospital’s smoke-free policy. Unfortunately, since discharge, Mr. D reports worsening auditory hallucinations despite adherence with his antipsychotic medication, clozapine, 600 mg at bedtime. Collateral information gathered from Mr. D’s mother confirms that he has been adherent with the discharge medication regimen; however, Mr. D has resumed smoking 1 pack of cigarettes daily. The treatment team suspects that his worsening psychosis is related to the decrease of blood clozapine level due to CYP induction by cigarette smoke.

Cotinine and smoking-related drug interactions

Vignette 1 illustrates the significant impact tobacco smoke can have on the effectiveness of a psychotropic medication. This is caused by polycyclic aromatic hydrocarbons induction of hepatic CYP1A2 isoenzymes. Clinicians should routinely screen patients for smoking status due to the potential for drug interactions. Common major CYP1A2 substrates include clozapine, olanzapine, duloxetine, and mirtazapine. Common minor CYP1A2 substrates include asenapine, chlorpromazine, haloperidol, perphenazine, ziprasidone, and zolpidem.²⁵ This case also highlights the potential advantage of collecting collateral information, and the utility of objective laboratory testing (such as cotinine monitoring) for optimal treatment because patient self-report can be unreliable and incomplete.^12,26

Vignette 2

Mr. B, age 34, has a history of cocaine use disorder and tobacco use disorder. He is referred to a treatment program and participates in a contingency management program for his substance use disorders. Biomarkers, including salivary cotinine, are used to assess Mr. B’s exposure to tobacco use. Mr. B and other participants in his program are eligible for prize draws if they are found to have samples that are negative for tobacco and other substances. There are other incentives in place for patients who show a reduced cotinine concentration.

Cotinine monitoring and contingency management

Clinicians can incorporate cotinine monitoring into existing SUD treatment. This is similar to the utilization of other biomarkers that are commonly used to identify recent illicit substance use or monitor adherence to treatment medications. For example, benzoylecgonine, a metabolite of cocaine, is frequently used to monitor abstinence from cocaine.²⁷ In addition, buprenorphine metabolites are used to monitor medication adherence and the possibility of diversion among patients being treated for opioid use disorder.²⁸ Smoking cessation may be associated with a decreased risk of relapse to illicit substances.²⁹

Treatments based on contingency management principles involve giving patients tangible rewards to reinforce desired (positive) behaviors. Smoking cessation can be confirmed by monitoring cotinine levels. Gayman et al⁹ found twice-weekly salivary testing was compatible with monitoring and promoting abstinence in a prize-based contingency management smoking cessation program. Most prior studies used urine cotinine measures to verify abstinence. Although highly reliable, urine samples require close monitoring to ensure sample validity, which can be a burden on staff and unpleasant for patients.⁹ It is also important to note that the rate of elimination of cotinine from saliva and urine are comparable. The half-life of cotinine is approximately 18 hours, and therefore the specificity of salivary test strips may be impacted during the first 4 to 5 days of abstinence. In the first few days of smoking cessation, a more intensive approach, such as quantifying urine cotinine levels and monitoring decline, may be appropriate.²³

Continue to: Vignette 3

Vignette 3

Ms. C, age 34 and pregnant, is admitted to an outpatient treatment program for alcohol use disorder. She also has generalized anxiety disorder and tobacco use disorder. In addition to attending group therapy sessions and self-reporting any recent alcohol consumption, Ms. C also undergoes alcohol breathalyzer tests and urine studies of alcohol metabolites to monitor abstinence from alcohol. She says that the regular laboratory screening for alcohol use gives her a sense of accountability and tangible evidence of change that positively impacts her treatment. When the treating psychiatrist recommends that Ms. C also consider addressing her tobacco use disorder, she asks if there is some way to include laboratory testing to monitor her smoking cessation.

Cotinine as a predictor of smoking status

Smoking abstinence rates during pregnancy are lower than that for other substances, and pregnant women may not be aware of the impact of smoking on fetal development.³⁰ Cotinine can be used to verify self-report of smoking status and severity.^10,31,32

Salivary cotinine tests are commercially available, relatively economical, and convenient to use when frequent monitoring is required.³² In general, based on established cut-off values that are unique to the specimen collected, the overall high specificity and sensitivity of salivary testing allows clinicians to predict smoker vs nonsmoker status with confidence. For example, a 2008 study reported a salivary cotinine cut-off level of 12 ng/mL for smokers.²¹ The sensitivity and specificity of this cut-off value for distinguishing cigarette smokers from never smokers were 96.7% and 96.9%, respectively.²¹

Additionally, some studies suggest that cotinine levels may be predictive of treatment outcomes and retention in SUD treatment programs.^33,34 One study of smoking cessation using nicotine replacement products found that compared with patients with lower baseline cotinine levels prior to treatment, patients with higher baseline cotinine plasma levels had lower smoking cessation success rates.³⁴

A few caveats

There are several limitations to quantitative measures of cotinine (Box 2^21,23). These include (but are not limited to) potential errors related to sample collection, storage, shipping, and analysis.²³ Compared with other methods, point-of-care cotinine measurement in saliva is noninvasive, simple, and requires less training to properly use.²³

Box 2

Continue to: Confirmation of smoking cessation...

Confirmation of smoking cessation can be monitored reliably within the clinical setting using cotinine monitoring. However, this is not a routine test, and there are no guidelines or consensus on how or when it should be used. The clinical feasibility of cotinine monitoring for psychiatric patients will depend on the cost of testing, methods used, amount of reimbursement for performing the tests, and how clinicians value such testing.³⁵

Bottom Line

Cotinine is a biomarker that can be used to detect tobacco use. Cotinine measurement can be used to monitor tobacco use and smoking cessation in psychiatric patients. Early detection and treatment of tobacco use disorder can improve patients’ health and reduce the incidence of acute and chronic illnesses. However, cotinine measurement is not a routine test, and there are no guidelines on how or when this test should be used.

Related Resources

• Peckham E, Brabyn S, Cook L, et al. Smoking cessation in severe mental ill health: what works? An updated systematic review and meta-analysis. BMC Psychiatry. 2017;17(1):252.
• Tidey JW, Miller ME. Smoking cessation and reduction in people with chronic mental illness. BMJ. 2015;351:h4065. doi: 10.1136/bmj.h4065

Drug Brand Names

Asenapine • Saphris
Buprenorphine • Sublocade
Clozapine • Clozaril
Duloxetine • Cymbalta
Haloperidol • Haldol
Mirtazapine • Remeron
Olanzapine • Zyprexa
Ziprasidone • Geodon
Zolpidem • Ambien

Cigarette smoking is common among patients with schizophrenia, mood disorders, anxiety disorders,^1-3 substance use disorders (SUDs),⁴ and other psychiatric disorders. Research suggests that compared with the general population, patients with SUDs consume more nicotine products and are more vulnerable to the effects of smoking.⁵ Despite the availability of effective treatments, many mental health professionals are reluctant to identify and treat tobacco use disorder,^6-8 or they prioritize other disorders over tobacco use. Early detection and treatment of tobacco use disorder can improve patients’ health and reduce the incidence of acute and chronic illness.

Cotinine is a biomarker that can be used to detect tobacco use. It can be measured in routine clinical practice by collecting urinary, serum, or salivary specimens, and used to monitor psychiatric patients’ tobacco use. Monitoring cotinine levels is similar to using other biomarkers to assess medication adherence or identify illicit substance use. A growing body of evidence supports the utility of cotinine screening as a part of a comprehensive substance use disorder treatment plan,^5,9,10 especially for:

• patients who have comorbid conditions that can be exacerbated by tobacco use, such as chronic obstructive pulmonary disease
• patients who are pregnant^11,12
• patients who are less reliable in self-report or who require objective testing for validation.

Routine clinical screening of tobacco use is recommended for all patients and early detection may facilitate earlier treatment. Several FDA-approved medications are available for smoking cessation¹³; however, discussion of treatment options is beyond the scope of this review. In this article, we describe how cotinine is measured and analyzed, 3 case vignettes that illustrate its potential clinical utility, and limitations to its use as a biomarker of tobacco use.

Methods of measuring cotinine

Cigarette smoking is associated with the absorption of nicotine, which is mainly metabolized by cytochrome P450 (CYP) 2A6 to 6 primary metabolites: cotinine, hydroxycotinine, norcotinine, nornicotine, cotinine oxide, and nicotine oxide.^14,15 Cotinine is the biomarker of choice for detecting use of tobacco/nicotine products due to its stability (it is not influenced by dietary or environmental factors), extended half-life (16 to 19 hours, compared with 2 hours for nicotine), and stable concentration throughout the day. Samples from saliva, urine, or blood can be analyzed through radioimmunoassay, enzyme-linked immunosorbent assay (ELISA), and gas/liquid chromatography.¹⁶ The specificity of cotinine for tobacco use is excellent, except for persons who are taking medications that contain nicotine.¹⁷

An advantage of cotinine over other biomarkers for smoking (such as carbon monoxide in expired air) is that the optimal cut-off points for cotinine are relatively uninfluenced by the prevalence of smoking in the population. The optimal cut-off levels used to detect current tobacco use may vary based on the sample or test used (saliva, urine, or plasma) and certain patient-specific factors (Box 1^11,16,18-21). However, for plasma or saliva cotinine, 16 ng/mL is the generally accepted cut-off level for detecting current tobacco use. A urinary cotinine cut-off level of 50 ng/mL is likely appropriate for most circumstances.¹⁷ Users of electronic nicotine delivery systems (electronic cigarettes) have been found to have cotinine levels similar to those of cigarette smokers.²²

Box 1

Assessment of cotinine levels in saliva may be considered for outpatient monitoring due to its noninvasive nature, tolerability, and the ability to collect multiple samples over a limited period.²³ Saliva cotinine levels correlate closely with blood concentrations. Urine cotinine levels offer some advantage because concentrations are 6 times higher in urine than in blood or saliva. For this reason, urine cotinine is the most widely used biomarker in individuals who use tobacco due to its high sensitivity, specificity, reliability, and noninvasive collection.²³ By using a lower urinary cut-off of ≥2.47 ng/mL, ELISA kits detect the highest sensitivity and specificity, which is useful for monitoring daily tobacco use.²⁴ This cut-off value was associated with 100% sensitivity and specificity, and these numbers declined with increases in the cut-off threshold.²³

Impact of cigarette smoking

The following 3 clinical vignettes illustrate the impact of tobacco use disorder on patients, and how cotinine might help with their treatment.

Continue to: Vignette 1

Mr. D, age 44, has a history of schizophrenia and has smoked 1 pack of cigarettes per day for the last 15 years. He was recently discharged from an inpatient psychiatric facility after his symptoms were stabilized. During his hospitalization, Mr. D used a nicotine-replacement product to comply with the hospital’s smoke-free policy. Unfortunately, since discharge, Mr. D reports worsening auditory hallucinations despite adherence with his antipsychotic medication, clozapine, 600 mg at bedtime. Collateral information gathered from Mr. D’s mother confirms that he has been adherent with the discharge medication regimen; however, Mr. D has resumed smoking 1 pack of cigarettes daily. The treatment team suspects that his worsening psychosis is related to the decrease of blood clozapine level due to CYP induction by cigarette smoke.

Cotinine and smoking-related drug interactions

Vignette 1 illustrates the significant impact tobacco smoke can have on the effectiveness of a psychotropic medication. This is caused by polycyclic aromatic hydrocarbons induction of hepatic CYP1A2 isoenzymes. Clinicians should routinely screen patients for smoking status due to the potential for drug interactions. Common major CYP1A2 substrates include clozapine, olanzapine, duloxetine, and mirtazapine. Common minor CYP1A2 substrates include asenapine, chlorpromazine, haloperidol, perphenazine, ziprasidone, and zolpidem.²⁵ This case also highlights the potential advantage of collecting collateral information, and the utility of objective laboratory testing (such as cotinine monitoring) for optimal treatment because patient self-report can be unreliable and incomplete.^12,26

Vignette 2

Mr. B, age 34, has a history of cocaine use disorder and tobacco use disorder. He is referred to a treatment program and participates in a contingency management program for his substance use disorders. Biomarkers, including salivary cotinine, are used to assess Mr. B’s exposure to tobacco use. Mr. B and other participants in his program are eligible for prize draws if they are found to have samples that are negative for tobacco and other substances. There are other incentives in place for patients who show a reduced cotinine concentration.

Cotinine monitoring and contingency management

Clinicians can incorporate cotinine monitoring into existing SUD treatment. This is similar to the utilization of other biomarkers that are commonly used to identify recent illicit substance use or monitor adherence to treatment medications. For example, benzoylecgonine, a metabolite of cocaine, is frequently used to monitor abstinence from cocaine.²⁷ In addition, buprenorphine metabolites are used to monitor medication adherence and the possibility of diversion among patients being treated for opioid use disorder.²⁸ Smoking cessation may be associated with a decreased risk of relapse to illicit substances.²⁹

Treatments based on contingency management principles involve giving patients tangible rewards to reinforce desired (positive) behaviors. Smoking cessation can be confirmed by monitoring cotinine levels. Gayman et al⁹ found twice-weekly salivary testing was compatible with monitoring and promoting abstinence in a prize-based contingency management smoking cessation program. Most prior studies used urine cotinine measures to verify abstinence. Although highly reliable, urine samples require close monitoring to ensure sample validity, which can be a burden on staff and unpleasant for patients.⁹ It is also important to note that the rate of elimination of cotinine from saliva and urine are comparable. The half-life of cotinine is approximately 18 hours, and therefore the specificity of salivary test strips may be impacted during the first 4 to 5 days of abstinence. In the first few days of smoking cessation, a more intensive approach, such as quantifying urine cotinine levels and monitoring decline, may be appropriate.²³

Continue to: Vignette 3

Vignette 3

Ms. C, age 34 and pregnant, is admitted to an outpatient treatment program for alcohol use disorder. She also has generalized anxiety disorder and tobacco use disorder. In addition to attending group therapy sessions and self-reporting any recent alcohol consumption, Ms. C also undergoes alcohol breathalyzer tests and urine studies of alcohol metabolites to monitor abstinence from alcohol. She says that the regular laboratory screening for alcohol use gives her a sense of accountability and tangible evidence of change that positively impacts her treatment. When the treating psychiatrist recommends that Ms. C also consider addressing her tobacco use disorder, she asks if there is some way to include laboratory testing to monitor her smoking cessation.

Cotinine as a predictor of smoking status

Smoking abstinence rates during pregnancy are lower than that for other substances, and pregnant women may not be aware of the impact of smoking on fetal development.³⁰ Cotinine can be used to verify self-report of smoking status and severity.^10,31,32

Salivary cotinine tests are commercially available, relatively economical, and convenient to use when frequent monitoring is required.³² In general, based on established cut-off values that are unique to the specimen collected, the overall high specificity and sensitivity of salivary testing allows clinicians to predict smoker vs nonsmoker status with confidence. For example, a 2008 study reported a salivary cotinine cut-off level of 12 ng/mL for smokers.²¹ The sensitivity and specificity of this cut-off value for distinguishing cigarette smokers from never smokers were 96.7% and 96.9%, respectively.²¹

Additionally, some studies suggest that cotinine levels may be predictive of treatment outcomes and retention in SUD treatment programs.^33,34 One study of smoking cessation using nicotine replacement products found that compared with patients with lower baseline cotinine levels prior to treatment, patients with higher baseline cotinine plasma levels had lower smoking cessation success rates.³⁴

A few caveats

There are several limitations to quantitative measures of cotinine (Box 2^21,23). These include (but are not limited to) potential errors related to sample collection, storage, shipping, and analysis.²³ Compared with other methods, point-of-care cotinine measurement in saliva is noninvasive, simple, and requires less training to properly use.²³

Box 2

Continue to: Confirmation of smoking cessation...

Confirmation of smoking cessation can be monitored reliably within the clinical setting using cotinine monitoring. However, this is not a routine test, and there are no guidelines or consensus on how or when it should be used. The clinical feasibility of cotinine monitoring for psychiatric patients will depend on the cost of testing, methods used, amount of reimbursement for performing the tests, and how clinicians value such testing.³⁵

Bottom Line

Cotinine is a biomarker that can be used to detect tobacco use. Cotinine measurement can be used to monitor tobacco use and smoking cessation in psychiatric patients. Early detection and treatment of tobacco use disorder can improve patients’ health and reduce the incidence of acute and chronic illnesses. However, cotinine measurement is not a routine test, and there are no guidelines on how or when this test should be used.

Related Resources

• Peckham E, Brabyn S, Cook L, et al. Smoking cessation in severe mental ill health: what works? An updated systematic review and meta-analysis. BMC Psychiatry. 2017;17(1):252.
• Tidey JW, Miller ME. Smoking cessation and reduction in people with chronic mental illness. BMJ. 2015;351:h4065. doi: 10.1136/bmj.h4065

Drug Brand Names

Asenapine • Saphris
Buprenorphine • Sublocade
Clozapine • Clozaril
Duloxetine • Cymbalta
Haloperidol • Haldol
Mirtazapine • Remeron
Olanzapine • Zyprexa
Ziprasidone • Geodon
Zolpidem • Ambien

Against the backdrop of an increasing opioid use epidemic and a marked acceleration of prescription opioid–related deaths,^1,2 there has been an impetus to explore the usefulness of alternative and co-analgesic agents to assist patients with chronic pain. Preclinical studies employing animal-based models of human pain syndromes have demonstrated that cannabis and chemicals derived from cannabis extracts may mitigate several pain conditions.³

Because there are significant comorbidities between psychiatric disorders and chronic pain, psychiatrists are likely to care for patients with chronic pain. As the availability of and interest in cannabinoid-based medications (CBM) increases, psychiatrists will need to be apprised of the utility, adverse effects, and potential drug interactions of these agents.

The endocannabinoid system and cannabis receptors

The endogenous cannabinoid (endocannabinoid) system is abundantly present within the peripheral and central nervous systems. The first identified, and best studied, endocannabinoids are N-arachidonoyl-ethanolamine (AEA; anandamide) and 2-arachidonoylglycerol (2-AG).⁴ Unlike typical neurotransmitters, AEA and 2-AG are not stored within vesicles within presynaptic neuron axons. Instead, they are lipophilic molecules produced on demand, synthesized from phospholipids (ie, arachidonic acid derivatives) at the membranes of post-synaptic neurons, and released into the synapse directly.⁵

Acting as retrograde messengers, the endocannabinoids traverse the synapse, binding to receptors located on the axons of the presynaptic neuron. Two receptors—CB1 and CB2—have been most extensively studied and characterized.^6,7 These receptors couple to Gi/o-proteins to inhibit adenylate cyclase, decreasing Ca2+ conductance and increasing K+ conductance.⁸ Once activated, cannabinoid receptors modulate neurotransmitter release from presynaptic axon terminals. Evidence points to a similar retrograde signaling between neurons and glial cells. Shortly after receptor activation, the endocannabinoids are deactivated by the actions of a transporter mechanism and enzyme degradation.^9,10

The endocannabinoid system and pain transmission

Cannabinoid receptors are present in pain transmission circuits spanning from the peripheral sensory nerve endings (from which pain signals originate) to the spinal cord and supraspinal regions within the brain.^11-14 CB1 receptors are abundantly present within the CNS, including regions involved in pain transmission. Binding to CB1 receptors, endocannabinoids modulate neurotransmission that impacts pain transmission centrally. Endocannabinoids can also indirectly modulate opiate and N-methyl-d-aspartate (NMDA) receptors involved in pain relay and transmission.¹⁵

By contrast, CB2 receptors are predominantly localized to peripheral tissues and immune cells, although there has been some discovery of their presence within the CNS (eg, on microglia). Endocannabinoid activation of CB2 receptors is thought to modulate the activity of peripheral afferent pain fibers and immune-mediated neuro­inflammatory processes—such as inhibition of prostaglandin synthesis and mast cell degranulation—that can precipitate and maintain chronic pain states.^16-18

Evidence garnered from preclinical (animal) studies points to the role of the endocannabinoid system in modulating normal pain transmission (see Manzanares et al³ for details). These studies offer a putative basis for understanding how exogenous cannabinoid congeners might serve to ameliorate pain transmission in pathophysiologic states, including chronic pain.

Continue to: Cannabinoid-based medications

Marijuana contains multiple components (cannabinoids). The most extensively studied are delta-9-tetrahydrocannabinol (THC) and cannabidiol (CBD). Because it predominantly binds CB1 receptors centrally, THC is the major psychoactive component of cannabis; it promotes sleep and appetite, influences anxiety, and produces the “high” associated with cannabis use. By contrast, CBD weakly binds CB1 and thus exerts minimal or no psycho­active effects.¹⁹

Cannabinoid absorption, metabolism, bioavailability, and clinical effects vary depending on the formulation and method of administration (Table 1).^20-22 THC and CBD content and potency in inhaled cannabis can vary significantly depending on the strains of the cannabis plant and manner of cultivation.²³ To standardize approaches for administering cannabinoids in clinical trials and for clinical use, researchers have developed pharmaceutical analogs that contain extracted chemicals or synthetic chemicals similar to THC and/or CBD.

In this article, CBM refers to smoked/vaporized herbal cannabis as well as pharmaceutical cannabis analogs. Table 2 summarizes the characteristics of CBM commonly used in studies investigating their use for managing pain conditions.

CBM for chronic pain

The literature base examining the role of CBM for managing chronic nonmalignant and malignant pain of varying etiologies is rapidly expanding. Randomized controlled trials (RCTs) have focused on inhaled/smoked products and related cannabinoid medications, some of which are FDA-approved (Table 2).

A multitude of other cannabinoid-based products are currently commercially available to consumers, including tincture and oil-based products; over-the-counter CBD products; and several other formulations of CBM (eg, edible and suppository products). Because such products are not standardized or quality-controlled,²⁴ RCTs have not assessed their efficacy for mitigating pain. Consequently, the findings summarized in this article do not address the utility of these agents.

Continue to: CBM for non-cancer pain

Neuropathic pain. Randomized controlled trials have assessed the pain-mitigating effects of various CBM, including inhaled cannabis, synthetic THC, plant-extracted CBD, and a THC/CBD spray. Studies have shown that inhaled/vaporized cannabis can produce short-term pain reduction in patients with chronic neuropathic pain of diverse etiologies, including diabetes mellitus-, HIV-, trauma-, and medication-induced neuropathies.^22,25,26 Similar beneficial effects have been observed with the use of cannabis analogues (eg, nabiximols).^25,26-29

Meta-analyses and systematic reviews have determined that most of these RCTs were of low-to-moderate quality.^26,30 Meta-analyses have revealed divergent and conflicting results because of differences in the inclusion and exclusion criteria used to select RCTs for analysis and differences in the standards with which the quality of evidence were determined.^25,30

Overall, the benefit of CBM for mitigating neuropathic pain is promising, but the effectiveness may not be robust.^30,31 Several noteworthy caveats limit the interpretation of the results of these RCTs:

• due to the small sample sizes and brief durations of study, questions remain regarding the extent to which effects are generalizable, whether the benefits are sustained, and whether adverse effects emerge over time with continued use
• most RCTs evaluated inhaled (herbal) cannabis and nabiximols; there is little data on the effectiveness of other CBM formulations^25,26,30
• the pain-mitigating effects of CBM were usually compared with those of placebo; the comparative efficacy against agents commonly used to treat neuropathic pain remains largely unexamined 

• these RCTs typically compared mean pain severity score differences between cannabis-treated and placebo groups using standard subjective rating scales of pain intensity, such as the Numerical Rating Scale or Visual Analogue Scale. Customarily, the pain literature has used a 30% or 50% reduction in pain severity from baseline as an indicator of significant clinical improvement.^32,33 The RCTs of CBM for neuropathic pain rarely used this standard, which makes it unclear whether CBM results in clinically significant pain reductions³⁰
• indirect measures of effectiveness (ie, whether using CBM reduces the need for opioids or other analgesics to manage pain) were seldom reported in these RCTs.

Due to these limitations, clinical guidelines and systematic reviews consider CBM as a third- or fourth-line therapy for patients experiencing chronic neuropathic pain for whom conventional agents such as anticonvulsants and antidepressants have failed.^34,35

Spasticity in multiple sclerosis (MS). Several RCTs have assessed the use of CBM for MS-related spasticity, although few were deemed to be high quality. Nabiximols and synthetic THC were effective in managing spasticity and reducing pain severity associated with muscle spasms.³⁶ Generally, investigations revealed that CBM were associated with improvements in subjective measures of spasticity, but these were not born out in clinical, objective measures.^26,37 The efficacy of smoked cannabis was uncertain.³⁷ The existing literature on CBM for MS-related spasticity does not address dosing, duration of effects, tolerability, or comparative effectiveness against conventional anti-spasm medications.^36,37

Continue to: Other chronic pain conditions

Other chronic pain conditions. CBM have also been studied for their usefulness in several other noncancer chronic conditions, including Crohn’s disease, inflammatory bowel disease, fibromyalgia, and other rheumatologic pain conditions.^22,31,38-40 However, a solid foundation of empirical work to inform their utility for managing pain in these conditions is lacking.

CBM for cancer pain

Anecdotal evidence suggests that inhaled cannabis has promising pain-mitigating effects in patients with advanced cancer.^41-43 There is a dearth of high-quality RCTs assessing the utility of CBM in patients with cancer pain.^43-45 The types of CBM used and dosing strategies varied across RCTs, which makes it difficult to infer how best to treat patients with cancer pain. The agents studied included nabiximols, THC spray, and synthetic THC capsules.^43-45 Although some studies have demonstrated that synthetic THC and nabiximols have potential for reducing subjective pain ratings compared with placebo,^46,47 these results were inconsistent.^46,48 Oromucosal nabiximols did not appear to confer any additional analgesic benefit in patients who were already prescribed opioids.^31,45

The benefit of CBM for mitigating cancer pain is promising, but it remains difficult to know how to position the use of CBM in managing cancer pain. Limitations in the cancer literature include:

• the RCTs addressing CBM use for cancer pain were often brief, which raises questions about the long-term effectiveness and adverse effects of these agents
• tolerability and dosing limits encountered due to adverse effects were seldom reported^43,45
• the types of cancer pain that patients had were often quite diverse. The small sample sizes and the heterogeneity of conditions included in these RCTs limit the ability to determine whether pain-mitigating effects might vary according to type of cancer-related pain.^31,45

Despite these limitations, some clinical guidelines and systematic reviews have suggested that CBM have some role in addressing refractory malignant pain conditions.⁴⁹

Psychiatric considerations related to CBM

As of November 2020, 36 states had legalized the use of cannabis for medical purposes, typically for painful conditions, despite the fact that empirical evidence to support their efficacy is mixed.⁵⁰ In light of recent changes in both the legal and popular attitudes regarding cannabis, the implications of legalizing CBM remains to be seen. For example, some research suggests that adults with pain are vulnerable to frequent nonmedical cannabis use and/or cannabis use disorder.⁵¹ Although well-intended, the legalization of CBM use might represent society’s next misstep in the quest to address the suffering of patients with chronic pain. Some evidence shows that cannabis use and cannabis use disorders increase in states that have legalized medical marijuana.^52,53 Psychiatrists will be on the front lines of addressing any potential consequences arising from the use of CBM for treating pain.

Continue to: Psychiatric disorders and CBM

Psychiatric disorders and CBM. The psychological impact of CBM use among patients enduring chronic pain can include sedation, cognitive/attention disturbance, and fatigue. These adverse effects can limit the utility of such agents.^22,29,45

Contraindications for CBM use, and conditions for which CBM ought to be used with caution, are listed in Table 3^54,55.The safety of CBM, particularly in patients with chronic pain and psychiatric disorders, has not been examined. Patients with psychiatric disorders may be poor candidates for medical cannabis. Epidemiologic data suggest that recreational cannabis use is positively associated both cross-sectionally and prospectively with psychotic spectrum disorders, depressive symptoms, and anxiety symptoms, including panic disorder.⁵⁶ Psychotic reactions have also been associated with CBM (dronabinol and nabilone).⁵⁷ Cannabis use also has been associated with an earlier onset of, and lower remission rates of, symptoms associated with bipolar disorder.^58,59 Consequently, patients who have been diagnosed with or are at risk for developing any of the aforementioned conditions may not be suitable candidates for CBM. If CBM are used, patients should be closely monitored for the emergence/exacerbation of psychiatric symptoms. The frequency and extent of follow-up is not clear, however. Because of its reduced propensity to produce psychoactive effects, CBD may be safer than THC for managing pain in individuals who have or are vulnerable to developing psychiatric disorders.

There is a lack of evidence to support the use of CBM for treating primary depressive disorders, general anxiety disorder, posttraumatic stress disorder, or psychosis.^60,61 Very low-quality evidence suggests that CBM could lead to a small improvement in anxiety among individuals with noncancer pain and MS.⁶⁰ However, interpreting causality is complicated. It is plausible that, for some patients, subjective improvement in pain severity may be related to reduced anxiety.⁶² Conversely, it is equally plausible that reductions in emotional distress may reduce the propensity to attend to, and thus magnify, pain severity. In the latter case, the indirect impact of reducing pain by modifying emotional distress can be impacted by the type and dose of CBM used. For example, low concentrations of THC produce anxiolytic effects, but high concentrations may be anxiety-provoking.^63,64

Several potential pharmacokinetic drug interactions may arise between herbal cannabis or CBM and other medications (Table 4^65,66). THC and CBD are both metabolized by cytochrome P450 (CYP) 2C19 and 3A4.^65,66 In addition, THC is also metabolized by CYP2C9. Medications that inhibit or induce these enzymes can increase or decrease the bioavailability of THC and CBD.⁶⁷

Simultaneously, cannabinoids can impact the bioavailability of co-prescribed medications (Table 5^66,68). Although such CYP enzyme interactions remain a theoretical possibility, it is uncertain whether significant perturbations in plasma concentrations (and clinical effects) have been encountered with prescription medications when co-administered with CBM.⁶⁹ Nonetheless, patients receiving CBM should be closely monitored for their response to prescribed medications.⁷⁰

Continue to: Potential CYP enzyme interactions...

Potential CYP enzyme interactions aside, clinicians need to consider the additive effects that may occur when CBM are combined with sympathomimetic agents (eg, tachycardia, hypertension); CNS depressants such as alcohol, benzodiazepines, and opioids (eg, drowsiness, ataxia); or anticholinergics (eg, tachycardia, confusion).⁷¹ Inhaled herbal cannabis contains mutagens and can result in lung damage, exacerbations of chronic bronchitis, and certain types of cancer.^54,72 Co-prescribing benzodiazepines may be contraindicated in light of their effects on respiratory rate and effort.

The THC contained in CBM produces hormonal effects (ie, significantly increases plasma levels of ghrelin and leptin and decreases peptide YY levels)⁷³ that affect appetite and can produce weight gain. This may be problematic for patients receiving psychoactive medications associated with increased risk of weight gain and dyslipidemia. Because of the association between cannabis use and motor vehicle accidents, patients whose jobs require them to drive or operate industrial equipment may not be ideal candidates for CBM, especially if such patients also consume alcohol or are prescribed benzodiazepines and/or sedative hypnotics.⁷⁴ Lastly, due to their lipophilicity, cannabinoids cross the placental barrier and can be found in breast milk⁷⁵ and therefore can affect pregnancy outcomes and neurodevelopment.

Bottom Line

The popularity of cannabinoid-based medications (CBM) for the treatment of chronic pain conditions is growing, but the interest in their use may be outpacing the evidence supporting their analgesic benefits. High-quality, well-controlled randomized controlled trials are needed to decipher whether, and to what extent, these agents can be positioned in chronic pain management. Because psychiatrists are likely to encounter patients considering, or receiving, CBM, they must be aware of the potential benefits, risks, and adverse effects of such treatments.

Related Resources

• Joshi KG. Cannabis-derived compounds: what you need to know. Current Psychiatry. 2020;19(10):64-65. doi:10.12788/ cp.0050
• Gupta S, Phalen T, Gupta S. Medical marijuana: do the benefits outweigh the risks? Current Psychiatry. 2018; 17(1):34-41.

Drug Brand Names

Ajulemic acid • Anabasum
Alprazolam • Xanax
Amitriptyline • Elavil
Aripiprazole • Abilify, Abilify Maintena
Buspirone • BuSpar
Cannabidiol • Epidiolex
Carbamazepine • Tegretol, Equetro
Cimetidine • Tagamet HB
Citalopram • Celexa
Clopidogrel • Plavix
Clozapine • Clozaril
Cyclosporine • Neoral, Sandimmune
Dronabinol • Marinol, Syndros
Duloxetine • Cymbalta
Fluoxetine • Prozac
Fluvoxamine • Luvox
Haloperidol • Haldol
Imipramine • Tofranil
Ketoconazole • Nizoral AD
Losartan • Cozaar
Midazolam • Versed
Mirtazapine • Remeron
Nabilone • Cesamet
Nabiximols • Sativex
Nefazodone • Serzone
Olanzapine • Zyprexa
Phenobarbital • Solfoton
Phenytoin • Dilantin
Ramelteon • Rozerem
Rifampin • Rifadin
Risperidone • Risperdal
Sertraline • Zoloft
Tamoxifen • Nolvadex
Topiramate • Topamax
Valproic acid • Depakote, Depakene
Venlafaxine • Effexor
Verapamil • Verelan
Zolpidem • Ambien

Against the backdrop of an increasing opioid use epidemic and a marked acceleration of prescription opioid–related deaths,^1,2 there has been an impetus to explore the usefulness of alternative and co-analgesic agents to assist patients with chronic pain. Preclinical studies employing animal-based models of human pain syndromes have demonstrated that cannabis and chemicals derived from cannabis extracts may mitigate several pain conditions.³

Because there are significant comorbidities between psychiatric disorders and chronic pain, psychiatrists are likely to care for patients with chronic pain. As the availability of and interest in cannabinoid-based medications (CBM) increases, psychiatrists will need to be apprised of the utility, adverse effects, and potential drug interactions of these agents.

The endocannabinoid system and cannabis receptors

The endogenous cannabinoid (endocannabinoid) system is abundantly present within the peripheral and central nervous systems. The first identified, and best studied, endocannabinoids are N-arachidonoyl-ethanolamine (AEA; anandamide) and 2-arachidonoylglycerol (2-AG).⁴ Unlike typical neurotransmitters, AEA and 2-AG are not stored within vesicles within presynaptic neuron axons. Instead, they are lipophilic molecules produced on demand, synthesized from phospholipids (ie, arachidonic acid derivatives) at the membranes of post-synaptic neurons, and released into the synapse directly.⁵

Acting as retrograde messengers, the endocannabinoids traverse the synapse, binding to receptors located on the axons of the presynaptic neuron. Two receptors—CB1 and CB2—have been most extensively studied and characterized.^6,7 These receptors couple to Gi/o-proteins to inhibit adenylate cyclase, decreasing Ca2+ conductance and increasing K+ conductance.⁸ Once activated, cannabinoid receptors modulate neurotransmitter release from presynaptic axon terminals. Evidence points to a similar retrograde signaling between neurons and glial cells. Shortly after receptor activation, the endocannabinoids are deactivated by the actions of a transporter mechanism and enzyme degradation.^9,10

The endocannabinoid system and pain transmission

Cannabinoid receptors are present in pain transmission circuits spanning from the peripheral sensory nerve endings (from which pain signals originate) to the spinal cord and supraspinal regions within the brain.^11-14 CB1 receptors are abundantly present within the CNS, including regions involved in pain transmission. Binding to CB1 receptors, endocannabinoids modulate neurotransmission that impacts pain transmission centrally. Endocannabinoids can also indirectly modulate opiate and N-methyl-d-aspartate (NMDA) receptors involved in pain relay and transmission.¹⁵

By contrast, CB2 receptors are predominantly localized to peripheral tissues and immune cells, although there has been some discovery of their presence within the CNS (eg, on microglia). Endocannabinoid activation of CB2 receptors is thought to modulate the activity of peripheral afferent pain fibers and immune-mediated neuro­inflammatory processes—such as inhibition of prostaglandin synthesis and mast cell degranulation—that can precipitate and maintain chronic pain states.^16-18

Evidence garnered from preclinical (animal) studies points to the role of the endocannabinoid system in modulating normal pain transmission (see Manzanares et al³ for details). These studies offer a putative basis for understanding how exogenous cannabinoid congeners might serve to ameliorate pain transmission in pathophysiologic states, including chronic pain.

Continue to: Cannabinoid-based medications

Marijuana contains multiple components (cannabinoids). The most extensively studied are delta-9-tetrahydrocannabinol (THC) and cannabidiol (CBD). Because it predominantly binds CB1 receptors centrally, THC is the major psychoactive component of cannabis; it promotes sleep and appetite, influences anxiety, and produces the “high” associated with cannabis use. By contrast, CBD weakly binds CB1 and thus exerts minimal or no psycho­active effects.¹⁹

Cannabinoid absorption, metabolism, bioavailability, and clinical effects vary depending on the formulation and method of administration (Table 1).^20-22 THC and CBD content and potency in inhaled cannabis can vary significantly depending on the strains of the cannabis plant and manner of cultivation.²³ To standardize approaches for administering cannabinoids in clinical trials and for clinical use, researchers have developed pharmaceutical analogs that contain extracted chemicals or synthetic chemicals similar to THC and/or CBD.

In this article, CBM refers to smoked/vaporized herbal cannabis as well as pharmaceutical cannabis analogs. Table 2 summarizes the characteristics of CBM commonly used in studies investigating their use for managing pain conditions.

CBM for chronic pain

The literature base examining the role of CBM for managing chronic nonmalignant and malignant pain of varying etiologies is rapidly expanding. Randomized controlled trials (RCTs) have focused on inhaled/smoked products and related cannabinoid medications, some of which are FDA-approved (Table 2).

A multitude of other cannabinoid-based products are currently commercially available to consumers, including tincture and oil-based products; over-the-counter CBD products; and several other formulations of CBM (eg, edible and suppository products). Because such products are not standardized or quality-controlled,²⁴ RCTs have not assessed their efficacy for mitigating pain. Consequently, the findings summarized in this article do not address the utility of these agents.

Continue to: CBM for non-cancer pain

Neuropathic pain. Randomized controlled trials have assessed the pain-mitigating effects of various CBM, including inhaled cannabis, synthetic THC, plant-extracted CBD, and a THC/CBD spray. Studies have shown that inhaled/vaporized cannabis can produce short-term pain reduction in patients with chronic neuropathic pain of diverse etiologies, including diabetes mellitus-, HIV-, trauma-, and medication-induced neuropathies.^22,25,26 Similar beneficial effects have been observed with the use of cannabis analogues (eg, nabiximols).^25,26-29

Meta-analyses and systematic reviews have determined that most of these RCTs were of low-to-moderate quality.^26,30 Meta-analyses have revealed divergent and conflicting results because of differences in the inclusion and exclusion criteria used to select RCTs for analysis and differences in the standards with which the quality of evidence were determined.^25,30

Overall, the benefit of CBM for mitigating neuropathic pain is promising, but the effectiveness may not be robust.^30,31 Several noteworthy caveats limit the interpretation of the results of these RCTs:

• due to the small sample sizes and brief durations of study, questions remain regarding the extent to which effects are generalizable, whether the benefits are sustained, and whether adverse effects emerge over time with continued use
• most RCTs evaluated inhaled (herbal) cannabis and nabiximols; there is little data on the effectiveness of other CBM formulations^25,26,30
• the pain-mitigating effects of CBM were usually compared with those of placebo; the comparative efficacy against agents commonly used to treat neuropathic pain remains largely unexamined 

• these RCTs typically compared mean pain severity score differences between cannabis-treated and placebo groups using standard subjective rating scales of pain intensity, such as the Numerical Rating Scale or Visual Analogue Scale. Customarily, the pain literature has used a 30% or 50% reduction in pain severity from baseline as an indicator of significant clinical improvement.^32,33 The RCTs of CBM for neuropathic pain rarely used this standard, which makes it unclear whether CBM results in clinically significant pain reductions³⁰
• indirect measures of effectiveness (ie, whether using CBM reduces the need for opioids or other analgesics to manage pain) were seldom reported in these RCTs.

Due to these limitations, clinical guidelines and systematic reviews consider CBM as a third- or fourth-line therapy for patients experiencing chronic neuropathic pain for whom conventional agents such as anticonvulsants and antidepressants have failed.^34,35

Spasticity in multiple sclerosis (MS). Several RCTs have assessed the use of CBM for MS-related spasticity, although few were deemed to be high quality. Nabiximols and synthetic THC were effective in managing spasticity and reducing pain severity associated with muscle spasms.³⁶ Generally, investigations revealed that CBM were associated with improvements in subjective measures of spasticity, but these were not born out in clinical, objective measures.^26,37 The efficacy of smoked cannabis was uncertain.³⁷ The existing literature on CBM for MS-related spasticity does not address dosing, duration of effects, tolerability, or comparative effectiveness against conventional anti-spasm medications.^36,37

Continue to: Other chronic pain conditions

Other chronic pain conditions. CBM have also been studied for their usefulness in several other noncancer chronic conditions, including Crohn’s disease, inflammatory bowel disease, fibromyalgia, and other rheumatologic pain conditions.^22,31,38-40 However, a solid foundation of empirical work to inform their utility for managing pain in these conditions is lacking.

CBM for cancer pain

Anecdotal evidence suggests that inhaled cannabis has promising pain-mitigating effects in patients with advanced cancer.^41-43 There is a dearth of high-quality RCTs assessing the utility of CBM in patients with cancer pain.^43-45 The types of CBM used and dosing strategies varied across RCTs, which makes it difficult to infer how best to treat patients with cancer pain. The agents studied included nabiximols, THC spray, and synthetic THC capsules.^43-45 Although some studies have demonstrated that synthetic THC and nabiximols have potential for reducing subjective pain ratings compared with placebo,^46,47 these results were inconsistent.^46,48 Oromucosal nabiximols did not appear to confer any additional analgesic benefit in patients who were already prescribed opioids.^31,45

The benefit of CBM for mitigating cancer pain is promising, but it remains difficult to know how to position the use of CBM in managing cancer pain. Limitations in the cancer literature include:

• the RCTs addressing CBM use for cancer pain were often brief, which raises questions about the long-term effectiveness and adverse effects of these agents
• tolerability and dosing limits encountered due to adverse effects were seldom reported^43,45
• the types of cancer pain that patients had were often quite diverse. The small sample sizes and the heterogeneity of conditions included in these RCTs limit the ability to determine whether pain-mitigating effects might vary according to type of cancer-related pain.^31,45

Despite these limitations, some clinical guidelines and systematic reviews have suggested that CBM have some role in addressing refractory malignant pain conditions.⁴⁹

Psychiatric considerations related to CBM

As of November 2020, 36 states had legalized the use of cannabis for medical purposes, typically for painful conditions, despite the fact that empirical evidence to support their efficacy is mixed.⁵⁰ In light of recent changes in both the legal and popular attitudes regarding cannabis, the implications of legalizing CBM remains to be seen. For example, some research suggests that adults with pain are vulnerable to frequent nonmedical cannabis use and/or cannabis use disorder.⁵¹ Although well-intended, the legalization of CBM use might represent society’s next misstep in the quest to address the suffering of patients with chronic pain. Some evidence shows that cannabis use and cannabis use disorders increase in states that have legalized medical marijuana.^52,53 Psychiatrists will be on the front lines of addressing any potential consequences arising from the use of CBM for treating pain.

Continue to: Psychiatric disorders and CBM

Psychiatric disorders and CBM. The psychological impact of CBM use among patients enduring chronic pain can include sedation, cognitive/attention disturbance, and fatigue. These adverse effects can limit the utility of such agents.^22,29,45

Contraindications for CBM use, and conditions for which CBM ought to be used with caution, are listed in Table 3^54,55.The safety of CBM, particularly in patients with chronic pain and psychiatric disorders, has not been examined. Patients with psychiatric disorders may be poor candidates for medical cannabis. Epidemiologic data suggest that recreational cannabis use is positively associated both cross-sectionally and prospectively with psychotic spectrum disorders, depressive symptoms, and anxiety symptoms, including panic disorder.⁵⁶ Psychotic reactions have also been associated with CBM (dronabinol and nabilone).⁵⁷ Cannabis use also has been associated with an earlier onset of, and lower remission rates of, symptoms associated with bipolar disorder.^58,59 Consequently, patients who have been diagnosed with or are at risk for developing any of the aforementioned conditions may not be suitable candidates for CBM. If CBM are used, patients should be closely monitored for the emergence/exacerbation of psychiatric symptoms. The frequency and extent of follow-up is not clear, however. Because of its reduced propensity to produce psychoactive effects, CBD may be safer than THC for managing pain in individuals who have or are vulnerable to developing psychiatric disorders.

There is a lack of evidence to support the use of CBM for treating primary depressive disorders, general anxiety disorder, posttraumatic stress disorder, or psychosis.^60,61 Very low-quality evidence suggests that CBM could lead to a small improvement in anxiety among individuals with noncancer pain and MS.⁶⁰ However, interpreting causality is complicated. It is plausible that, for some patients, subjective improvement in pain severity may be related to reduced anxiety.⁶² Conversely, it is equally plausible that reductions in emotional distress may reduce the propensity to attend to, and thus magnify, pain severity. In the latter case, the indirect impact of reducing pain by modifying emotional distress can be impacted by the type and dose of CBM used. For example, low concentrations of THC produce anxiolytic effects, but high concentrations may be anxiety-provoking.^63,64

Several potential pharmacokinetic drug interactions may arise between herbal cannabis or CBM and other medications (Table 4^65,66). THC and CBD are both metabolized by cytochrome P450 (CYP) 2C19 and 3A4.^65,66 In addition, THC is also metabolized by CYP2C9. Medications that inhibit or induce these enzymes can increase or decrease the bioavailability of THC and CBD.⁶⁷

Simultaneously, cannabinoids can impact the bioavailability of co-prescribed medications (Table 5^66,68). Although such CYP enzyme interactions remain a theoretical possibility, it is uncertain whether significant perturbations in plasma concentrations (and clinical effects) have been encountered with prescription medications when co-administered with CBM.⁶⁹ Nonetheless, patients receiving CBM should be closely monitored for their response to prescribed medications.⁷⁰

Continue to: Potential CYP enzyme interactions...

Potential CYP enzyme interactions aside, clinicians need to consider the additive effects that may occur when CBM are combined with sympathomimetic agents (eg, tachycardia, hypertension); CNS depressants such as alcohol, benzodiazepines, and opioids (eg, drowsiness, ataxia); or anticholinergics (eg, tachycardia, confusion).⁷¹ Inhaled herbal cannabis contains mutagens and can result in lung damage, exacerbations of chronic bronchitis, and certain types of cancer.^54,72 Co-prescribing benzodiazepines may be contraindicated in light of their effects on respiratory rate and effort.

The THC contained in CBM produces hormonal effects (ie, significantly increases plasma levels of ghrelin and leptin and decreases peptide YY levels)⁷³ that affect appetite and can produce weight gain. This may be problematic for patients receiving psychoactive medications associated with increased risk of weight gain and dyslipidemia. Because of the association between cannabis use and motor vehicle accidents, patients whose jobs require them to drive or operate industrial equipment may not be ideal candidates for CBM, especially if such patients also consume alcohol or are prescribed benzodiazepines and/or sedative hypnotics.⁷⁴ Lastly, due to their lipophilicity, cannabinoids cross the placental barrier and can be found in breast milk⁷⁵ and therefore can affect pregnancy outcomes and neurodevelopment.

Bottom Line

The popularity of cannabinoid-based medications (CBM) for the treatment of chronic pain conditions is growing, but the interest in their use may be outpacing the evidence supporting their analgesic benefits. High-quality, well-controlled randomized controlled trials are needed to decipher whether, and to what extent, these agents can be positioned in chronic pain management. Because psychiatrists are likely to encounter patients considering, or receiving, CBM, they must be aware of the potential benefits, risks, and adverse effects of such treatments.

Related Resources

• Joshi KG. Cannabis-derived compounds: what you need to know. Current Psychiatry. 2020;19(10):64-65. doi:10.12788/ cp.0050
• Gupta S, Phalen T, Gupta S. Medical marijuana: do the benefits outweigh the risks? Current Psychiatry. 2018; 17(1):34-41.

Drug Brand Names

Ajulemic acid • Anabasum
Alprazolam • Xanax
Amitriptyline • Elavil
Aripiprazole • Abilify, Abilify Maintena
Buspirone • BuSpar
Cannabidiol • Epidiolex
Carbamazepine • Tegretol, Equetro
Cimetidine • Tagamet HB
Citalopram • Celexa
Clopidogrel • Plavix
Clozapine • Clozaril
Cyclosporine • Neoral, Sandimmune
Dronabinol • Marinol, Syndros
Duloxetine • Cymbalta
Fluoxetine • Prozac
Fluvoxamine • Luvox
Haloperidol • Haldol
Imipramine • Tofranil
Ketoconazole • Nizoral AD
Losartan • Cozaar
Midazolam • Versed
Mirtazapine • Remeron
Nabilone • Cesamet
Nabiximols • Sativex
Nefazodone • Serzone
Olanzapine • Zyprexa
Phenobarbital • Solfoton
Phenytoin • Dilantin
Ramelteon • Rozerem
Rifampin • Rifadin
Risperidone • Risperdal
Sertraline • Zoloft
Tamoxifen • Nolvadex
Topiramate • Topamax
Valproic acid • Depakote, Depakene
Venlafaxine • Effexor
Verapamil • Verelan
Zolpidem • Ambien

Against the backdrop of an increasing opioid use epidemic and a marked acceleration of prescription opioid–related deaths,^1,2 there has been an impetus to explore the usefulness of alternative and co-analgesic agents to assist patients with chronic pain. Preclinical studies employing animal-based models of human pain syndromes have demonstrated that cannabis and chemicals derived from cannabis extracts may mitigate several pain conditions.³

Because there are significant comorbidities between psychiatric disorders and chronic pain, psychiatrists are likely to care for patients with chronic pain. As the availability of and interest in cannabinoid-based medications (CBM) increases, psychiatrists will need to be apprised of the utility, adverse effects, and potential drug interactions of these agents.

The endocannabinoid system and cannabis receptors

The endogenous cannabinoid (endocannabinoid) system is abundantly present within the peripheral and central nervous systems. The first identified, and best studied, endocannabinoids are N-arachidonoyl-ethanolamine (AEA; anandamide) and 2-arachidonoylglycerol (2-AG).⁴ Unlike typical neurotransmitters, AEA and 2-AG are not stored within vesicles within presynaptic neuron axons. Instead, they are lipophilic molecules produced on demand, synthesized from phospholipids (ie, arachidonic acid derivatives) at the membranes of post-synaptic neurons, and released into the synapse directly.⁵

Acting as retrograde messengers, the endocannabinoids traverse the synapse, binding to receptors located on the axons of the presynaptic neuron. Two receptors—CB1 and CB2—have been most extensively studied and characterized.^6,7 These receptors couple to Gi/o-proteins to inhibit adenylate cyclase, decreasing Ca2+ conductance and increasing K+ conductance.⁸ Once activated, cannabinoid receptors modulate neurotransmitter release from presynaptic axon terminals. Evidence points to a similar retrograde signaling between neurons and glial cells. Shortly after receptor activation, the endocannabinoids are deactivated by the actions of a transporter mechanism and enzyme degradation.^9,10

The endocannabinoid system and pain transmission

Cannabinoid receptors are present in pain transmission circuits spanning from the peripheral sensory nerve endings (from which pain signals originate) to the spinal cord and supraspinal regions within the brain.^11-14 CB1 receptors are abundantly present within the CNS, including regions involved in pain transmission. Binding to CB1 receptors, endocannabinoids modulate neurotransmission that impacts pain transmission centrally. Endocannabinoids can also indirectly modulate opiate and N-methyl-d-aspartate (NMDA) receptors involved in pain relay and transmission.¹⁵

By contrast, CB2 receptors are predominantly localized to peripheral tissues and immune cells, although there has been some discovery of their presence within the CNS (eg, on microglia). Endocannabinoid activation of CB2 receptors is thought to modulate the activity of peripheral afferent pain fibers and immune-mediated neuro­inflammatory processes—such as inhibition of prostaglandin synthesis and mast cell degranulation—that can precipitate and maintain chronic pain states.^16-18

Evidence garnered from preclinical (animal) studies points to the role of the endocannabinoid system in modulating normal pain transmission (see Manzanares et al³ for details). These studies offer a putative basis for understanding how exogenous cannabinoid congeners might serve to ameliorate pain transmission in pathophysiologic states, including chronic pain.

Continue to: Cannabinoid-based medications

Marijuana contains multiple components (cannabinoids). The most extensively studied are delta-9-tetrahydrocannabinol (THC) and cannabidiol (CBD). Because it predominantly binds CB1 receptors centrally, THC is the major psychoactive component of cannabis; it promotes sleep and appetite, influences anxiety, and produces the “high” associated with cannabis use. By contrast, CBD weakly binds CB1 and thus exerts minimal or no psycho­active effects.¹⁹

Cannabinoid absorption, metabolism, bioavailability, and clinical effects vary depending on the formulation and method of administration (Table 1).^20-22 THC and CBD content and potency in inhaled cannabis can vary significantly depending on the strains of the cannabis plant and manner of cultivation.²³ To standardize approaches for administering cannabinoids in clinical trials and for clinical use, researchers have developed pharmaceutical analogs that contain extracted chemicals or synthetic chemicals similar to THC and/or CBD.

In this article, CBM refers to smoked/vaporized herbal cannabis as well as pharmaceutical cannabis analogs. Table 2 summarizes the characteristics of CBM commonly used in studies investigating their use for managing pain conditions.

CBM for chronic pain

The literature base examining the role of CBM for managing chronic nonmalignant and malignant pain of varying etiologies is rapidly expanding. Randomized controlled trials (RCTs) have focused on inhaled/smoked products and related cannabinoid medications, some of which are FDA-approved (Table 2).

A multitude of other cannabinoid-based products are currently commercially available to consumers, including tincture and oil-based products; over-the-counter CBD products; and several other formulations of CBM (eg, edible and suppository products). Because such products are not standardized or quality-controlled,²⁴ RCTs have not assessed their efficacy for mitigating pain. Consequently, the findings summarized in this article do not address the utility of these agents.

Continue to: CBM for non-cancer pain

Neuropathic pain. Randomized controlled trials have assessed the pain-mitigating effects of various CBM, including inhaled cannabis, synthetic THC, plant-extracted CBD, and a THC/CBD spray. Studies have shown that inhaled/vaporized cannabis can produce short-term pain reduction in patients with chronic neuropathic pain of diverse etiologies, including diabetes mellitus-, HIV-, trauma-, and medication-induced neuropathies.^22,25,26 Similar beneficial effects have been observed with the use of cannabis analogues (eg, nabiximols).^25,26-29

Meta-analyses and systematic reviews have determined that most of these RCTs were of low-to-moderate quality.^26,30 Meta-analyses have revealed divergent and conflicting results because of differences in the inclusion and exclusion criteria used to select RCTs for analysis and differences in the standards with which the quality of evidence were determined.^25,30

Overall, the benefit of CBM for mitigating neuropathic pain is promising, but the effectiveness may not be robust.^30,31 Several noteworthy caveats limit the interpretation of the results of these RCTs:

• due to the small sample sizes and brief durations of study, questions remain regarding the extent to which effects are generalizable, whether the benefits are sustained, and whether adverse effects emerge over time with continued use
• most RCTs evaluated inhaled (herbal) cannabis and nabiximols; there is little data on the effectiveness of other CBM formulations^25,26,30
• the pain-mitigating effects of CBM were usually compared with those of placebo; the comparative efficacy against agents commonly used to treat neuropathic pain remains largely unexamined 

• these RCTs typically compared mean pain severity score differences between cannabis-treated and placebo groups using standard subjective rating scales of pain intensity, such as the Numerical Rating Scale or Visual Analogue Scale. Customarily, the pain literature has used a 30% or 50% reduction in pain severity from baseline as an indicator of significant clinical improvement.^32,33 The RCTs of CBM for neuropathic pain rarely used this standard, which makes it unclear whether CBM results in clinically significant pain reductions³⁰
• indirect measures of effectiveness (ie, whether using CBM reduces the need for opioids or other analgesics to manage pain) were seldom reported in these RCTs.

Due to these limitations, clinical guidelines and systematic reviews consider CBM as a third- or fourth-line therapy for patients experiencing chronic neuropathic pain for whom conventional agents such as anticonvulsants and antidepressants have failed.^34,35

Spasticity in multiple sclerosis (MS). Several RCTs have assessed the use of CBM for MS-related spasticity, although few were deemed to be high quality. Nabiximols and synthetic THC were effective in managing spasticity and reducing pain severity associated with muscle spasms.³⁶ Generally, investigations revealed that CBM were associated with improvements in subjective measures of spasticity, but these were not born out in clinical, objective measures.^26,37 The efficacy of smoked cannabis was uncertain.³⁷ The existing literature on CBM for MS-related spasticity does not address dosing, duration of effects, tolerability, or comparative effectiveness against conventional anti-spasm medications.^36,37

Continue to: Other chronic pain conditions

Other chronic pain conditions. CBM have also been studied for their usefulness in several other noncancer chronic conditions, including Crohn’s disease, inflammatory bowel disease, fibromyalgia, and other rheumatologic pain conditions.^22,31,38-40 However, a solid foundation of empirical work to inform their utility for managing pain in these conditions is lacking.

CBM for cancer pain

Anecdotal evidence suggests that inhaled cannabis has promising pain-mitigating effects in patients with advanced cancer.^41-43 There is a dearth of high-quality RCTs assessing the utility of CBM in patients with cancer pain.^43-45 The types of CBM used and dosing strategies varied across RCTs, which makes it difficult to infer how best to treat patients with cancer pain. The agents studied included nabiximols, THC spray, and synthetic THC capsules.^43-45 Although some studies have demonstrated that synthetic THC and nabiximols have potential for reducing subjective pain ratings compared with placebo,^46,47 these results were inconsistent.^46,48 Oromucosal nabiximols did not appear to confer any additional analgesic benefit in patients who were already prescribed opioids.^31,45

The benefit of CBM for mitigating cancer pain is promising, but it remains difficult to know how to position the use of CBM in managing cancer pain. Limitations in the cancer literature include:

• the RCTs addressing CBM use for cancer pain were often brief, which raises questions about the long-term effectiveness and adverse effects of these agents
• tolerability and dosing limits encountered due to adverse effects were seldom reported^43,45
• the types of cancer pain that patients had were often quite diverse. The small sample sizes and the heterogeneity of conditions included in these RCTs limit the ability to determine whether pain-mitigating effects might vary according to type of cancer-related pain.^31,45

Despite these limitations, some clinical guidelines and systematic reviews have suggested that CBM have some role in addressing refractory malignant pain conditions.⁴⁹

Psychiatric considerations related to CBM

As of November 2020, 36 states had legalized the use of cannabis for medical purposes, typically for painful conditions, despite the fact that empirical evidence to support their efficacy is mixed.⁵⁰ In light of recent changes in both the legal and popular attitudes regarding cannabis, the implications of legalizing CBM remains to be seen. For example, some research suggests that adults with pain are vulnerable to frequent nonmedical cannabis use and/or cannabis use disorder.⁵¹ Although well-intended, the legalization of CBM use might represent society’s next misstep in the quest to address the suffering of patients with chronic pain. Some evidence shows that cannabis use and cannabis use disorders increase in states that have legalized medical marijuana.^52,53 Psychiatrists will be on the front lines of addressing any potential consequences arising from the use of CBM for treating pain.

Continue to: Psychiatric disorders and CBM

Psychiatric disorders and CBM. The psychological impact of CBM use among patients enduring chronic pain can include sedation, cognitive/attention disturbance, and fatigue. These adverse effects can limit the utility of such agents.^22,29,45

Contraindications for CBM use, and conditions for which CBM ought to be used with caution, are listed in Table 3^54,55.The safety of CBM, particularly in patients with chronic pain and psychiatric disorders, has not been examined. Patients with psychiatric disorders may be poor candidates for medical cannabis. Epidemiologic data suggest that recreational cannabis use is positively associated both cross-sectionally and prospectively with psychotic spectrum disorders, depressive symptoms, and anxiety symptoms, including panic disorder.⁵⁶ Psychotic reactions have also been associated with CBM (dronabinol and nabilone).⁵⁷ Cannabis use also has been associated with an earlier onset of, and lower remission rates of, symptoms associated with bipolar disorder.^58,59 Consequently, patients who have been diagnosed with or are at risk for developing any of the aforementioned conditions may not be suitable candidates for CBM. If CBM are used, patients should be closely monitored for the emergence/exacerbation of psychiatric symptoms. The frequency and extent of follow-up is not clear, however. Because of its reduced propensity to produce psychoactive effects, CBD may be safer than THC for managing pain in individuals who have or are vulnerable to developing psychiatric disorders.

There is a lack of evidence to support the use of CBM for treating primary depressive disorders, general anxiety disorder, posttraumatic stress disorder, or psychosis.^60,61 Very low-quality evidence suggests that CBM could lead to a small improvement in anxiety among individuals with noncancer pain and MS.⁶⁰ However, interpreting causality is complicated. It is plausible that, for some patients, subjective improvement in pain severity may be related to reduced anxiety.⁶² Conversely, it is equally plausible that reductions in emotional distress may reduce the propensity to attend to, and thus magnify, pain severity. In the latter case, the indirect impact of reducing pain by modifying emotional distress can be impacted by the type and dose of CBM used. For example, low concentrations of THC produce anxiolytic effects, but high concentrations may be anxiety-provoking.^63,64

Several potential pharmacokinetic drug interactions may arise between herbal cannabis or CBM and other medications (Table 4^65,66). THC and CBD are both metabolized by cytochrome P450 (CYP) 2C19 and 3A4.^65,66 In addition, THC is also metabolized by CYP2C9. Medications that inhibit or induce these enzymes can increase or decrease the bioavailability of THC and CBD.⁶⁷

Simultaneously, cannabinoids can impact the bioavailability of co-prescribed medications (Table 5^66,68). Although such CYP enzyme interactions remain a theoretical possibility, it is uncertain whether significant perturbations in plasma concentrations (and clinical effects) have been encountered with prescription medications when co-administered with CBM.⁶⁹ Nonetheless, patients receiving CBM should be closely monitored for their response to prescribed medications.⁷⁰

Continue to: Potential CYP enzyme interactions...

Potential CYP enzyme interactions aside, clinicians need to consider the additive effects that may occur when CBM are combined with sympathomimetic agents (eg, tachycardia, hypertension); CNS depressants such as alcohol, benzodiazepines, and opioids (eg, drowsiness, ataxia); or anticholinergics (eg, tachycardia, confusion).⁷¹ Inhaled herbal cannabis contains mutagens and can result in lung damage, exacerbations of chronic bronchitis, and certain types of cancer.^54,72 Co-prescribing benzodiazepines may be contraindicated in light of their effects on respiratory rate and effort.

The THC contained in CBM produces hormonal effects (ie, significantly increases plasma levels of ghrelin and leptin and decreases peptide YY levels)⁷³ that affect appetite and can produce weight gain. This may be problematic for patients receiving psychoactive medications associated with increased risk of weight gain and dyslipidemia. Because of the association between cannabis use and motor vehicle accidents, patients whose jobs require them to drive or operate industrial equipment may not be ideal candidates for CBM, especially if such patients also consume alcohol or are prescribed benzodiazepines and/or sedative hypnotics.⁷⁴ Lastly, due to their lipophilicity, cannabinoids cross the placental barrier and can be found in breast milk⁷⁵ and therefore can affect pregnancy outcomes and neurodevelopment.

Bottom Line

The popularity of cannabinoid-based medications (CBM) for the treatment of chronic pain conditions is growing, but the interest in their use may be outpacing the evidence supporting their analgesic benefits. High-quality, well-controlled randomized controlled trials are needed to decipher whether, and to what extent, these agents can be positioned in chronic pain management. Because psychiatrists are likely to encounter patients considering, or receiving, CBM, they must be aware of the potential benefits, risks, and adverse effects of such treatments.

Related Resources

• Joshi KG. Cannabis-derived compounds: what you need to know. Current Psychiatry. 2020;19(10):64-65. doi:10.12788/ cp.0050
• Gupta S, Phalen T, Gupta S. Medical marijuana: do the benefits outweigh the risks? Current Psychiatry. 2018; 17(1):34-41.

Drug Brand Names

Ajulemic acid • Anabasum
Alprazolam • Xanax
Amitriptyline • Elavil
Aripiprazole • Abilify, Abilify Maintena
Buspirone • BuSpar
Cannabidiol • Epidiolex
Carbamazepine • Tegretol, Equetro
Cimetidine • Tagamet HB
Citalopram • Celexa
Clopidogrel • Plavix
Clozapine • Clozaril
Cyclosporine • Neoral, Sandimmune
Dronabinol • Marinol, Syndros
Duloxetine • Cymbalta
Fluoxetine • Prozac
Fluvoxamine • Luvox
Haloperidol • Haldol
Imipramine • Tofranil
Ketoconazole • Nizoral AD
Losartan • Cozaar
Midazolam • Versed
Mirtazapine • Remeron
Nabilone • Cesamet
Nabiximols • Sativex
Nefazodone • Serzone
Olanzapine • Zyprexa
Phenobarbital • Solfoton
Phenytoin • Dilantin
Ramelteon • Rozerem
Rifampin • Rifadin
Risperidone • Risperdal
Sertraline • Zoloft
Tamoxifen • Nolvadex
Topiramate • Topamax
Valproic acid • Depakote, Depakene
Venlafaxine • Effexor
Verapamil • Verelan
Zolpidem • Ambien

Emerging data demonstrate that untreated perinatal anxiety is associated with negative outcomes, including an increased risk for suicide.¹ A 2017 systematic review and meta-analysis that included 102 studies with a total of 221,974 women from 34 countries found that the prevalence of self-reported anxiety symptoms and any anxiety disorder was 22.9% and 15.2%, respectively, across the 3 trimesters.¹ During pregnancy, anxiety disorders (eg, generalized anxiety disorder) and anxiety-related disorders (eg, obsessive-compulsive disorder [OCD] and posttraumatic stress disorder [PTSD]) can present as new illnesses or as a reoccurrence of an existing illness. Patients with pre-existing OCD may notice that the nature of their obsessions is changing. Women with pre-existing PTSD may have their symptoms triggered by pregnancy or delivery or may develop PTSD as a result of a traumatic delivery. Anxiety is frequently comorbid with depression, and high anxiety during pregnancy is one of the strongest risk factors for depression.^1,2

In light of this data, awareness and recognition of perinatal anxiety is critical. In this article, we describe how to accurately assess perinatal anxiety by avoiding assumptions and asking key questions during the clinical interview.

Avoid these common assumptions

Assessment begins with avoiding assumptions typically associated with maternal mental health. One common assumption is that pregnancy is a joyous occasion for all women. Pregnancy can be a stressful time that has its own unique difficulties, including the potential to develop or have a relapse of a mental illness. Another assumption is that the only concern is “postpartum depression.” In actuality, a significant percentage of women will experience depression during their pregnancy (not just in the postpartum period), and many other psychiatric illnesses are common during the perinatal period, including anxiety disorders.

Conduct a focused interview

Risk factors associated with antenatal anxiety include²:

• previous history of mental illness (particularly a history of anxiety and depression and a history of psychiatric treatment)
• lack of partner or social support
• history of abuse or domestic violence
• unplanned or unwanted pregnancy
• adverse events in life and high perceived stress
• present/past pregnancy complications
• pregnancy loss.

Symptoms of anxiety. The presence of anxiety or worrying does not necessarily mean a mother has an anxiety disorder. Using the DSM-5 as a guide, we should use the questions outlined in the following sections to inquire about all of the symptoms related to a particular illness, the pervasiveness of these symptoms, and to what extent these symptoms impair a woman’s ability to function and carry out her usual activities.³

Past psychiatric history. Ask your patient the following: Have you previously experienced anxiety and/or depressive symptoms? Were those symptoms limited only to times when you were pregnant or postpartum? Were your symptoms severe enough to disrupt your life (job, school, relationships, ability to complete daily tasks)? What treatments were effective for your symptoms? What treatments were ineffective?³

Social factors. Learn more about your patient’s support systems by asking: Who do you consider to be part of your social support? How is your relationship with your social support? Are there challenges in your relationship with your friends, family, or partner? If yes, what are those challenges? Are there other children in the home, and do you have support for them? Is your home environment safe? Do you feel that you have what you need for the baby? What stressors are you currently experiencing? Do you attend support groups for expectant mothers? Are you engaged in perinatal care?³

Continue to: Given the high prevalence...

Given the high prevalence of interpersonal violence in women of reproductive age, all patients should be screened for this. The American College of Obstetricians and Gynecologists Committee on Health Care for Underserved Women recommends screening for interpersonal violence at the first visit during the perinatal period, during each trimester, and at the postpartum visit (at minimum).⁴ Potential screening questions include (but are not limited to): Have you and/or your children ever been threatened by or felt afraid of your partner? When you argue with your partner, do either of you get physical? Has your partner ever physically hurt you (eg, hit, choked)? Do you feel safe at home? Do you have a safe place to go with resources you and your children will need in case of an emergency?^4-6

Feelings toward pregnancy, past/current pregnancy complications, and pregnancy loss. Ask your patient: Was this pregnancy planned? How do you feel about your pregnancy? How do you see yourself as a mother? Do you currently have pregnancy complications and/or have had them in the past, and, if so, what are/were they? Have you lost a pregnancy? If so, what was that like? Do you have fears related to childbirth, and, if so, what are they?³

Intrusive thoughts about harming the baby. Intrusive thoughts are common in postpartum women with anxiety disorders, including OCD.⁷ Merely asking patients if they’ve had thoughts of harming their baby is incomplete and insufficient to assess for intrusive thoughts. This question does not distinguish between intrusive thoughts and homicidal ideation; this distinction is absolutely necessary given the difference in potential risk to the infant.

Intrusive thoughts are generally associated with a low risk of mothers acting on their thoughts. These thoughts are typically ego dystonic and, in the most severe form, can be distressing to an extent that they cause behavioral changes, such as avoiding bathing the infant, avoiding diaper changes, avoiding knives, or separating themselves from the infant.⁷ On the contrary, having homicidal ideation carries a higher risk for harm to the infant. Homicidal ideation may be seen in patients with co-occurring psychosis, poor reality testing, and delusions.^5,7

Questions such as “Do you worry about harm coming to your baby?” “Do you worry about you causing harm to your baby?” and “Have you had an upsetting thought about harming your baby?” are more likely to reveal intrusive thoughts and prompt further exploration. Statements such as “Some people tell me that they have distressing thoughts about harm coming to their baby” can gently open the door to a having a dialogue about such thoughts. This dialogue is significantly important in making informed assessments as we develop comprehensive treatment plans.

Emerging data demonstrate that untreated perinatal anxiety is associated with negative outcomes, including an increased risk for suicide.¹ A 2017 systematic review and meta-analysis that included 102 studies with a total of 221,974 women from 34 countries found that the prevalence of self-reported anxiety symptoms and any anxiety disorder was 22.9% and 15.2%, respectively, across the 3 trimesters.¹ During pregnancy, anxiety disorders (eg, generalized anxiety disorder) and anxiety-related disorders (eg, obsessive-compulsive disorder [OCD] and posttraumatic stress disorder [PTSD]) can present as new illnesses or as a reoccurrence of an existing illness. Patients with pre-existing OCD may notice that the nature of their obsessions is changing. Women with pre-existing PTSD may have their symptoms triggered by pregnancy or delivery or may develop PTSD as a result of a traumatic delivery. Anxiety is frequently comorbid with depression, and high anxiety during pregnancy is one of the strongest risk factors for depression.^1,2

In light of this data, awareness and recognition of perinatal anxiety is critical. In this article, we describe how to accurately assess perinatal anxiety by avoiding assumptions and asking key questions during the clinical interview.

Avoid these common assumptions

Assessment begins with avoiding assumptions typically associated with maternal mental health. One common assumption is that pregnancy is a joyous occasion for all women. Pregnancy can be a stressful time that has its own unique difficulties, including the potential to develop or have a relapse of a mental illness. Another assumption is that the only concern is “postpartum depression.” In actuality, a significant percentage of women will experience depression during their pregnancy (not just in the postpartum period), and many other psychiatric illnesses are common during the perinatal period, including anxiety disorders.

Conduct a focused interview

Risk factors associated with antenatal anxiety include²:

• previous history of mental illness (particularly a history of anxiety and depression and a history of psychiatric treatment)
• lack of partner or social support
• history of abuse or domestic violence
• unplanned or unwanted pregnancy
• adverse events in life and high perceived stress
• present/past pregnancy complications
• pregnancy loss.

Symptoms of anxiety. The presence of anxiety or worrying does not necessarily mean a mother has an anxiety disorder. Using the DSM-5 as a guide, we should use the questions outlined in the following sections to inquire about all of the symptoms related to a particular illness, the pervasiveness of these symptoms, and to what extent these symptoms impair a woman’s ability to function and carry out her usual activities.³

Past psychiatric history. Ask your patient the following: Have you previously experienced anxiety and/or depressive symptoms? Were those symptoms limited only to times when you were pregnant or postpartum? Were your symptoms severe enough to disrupt your life (job, school, relationships, ability to complete daily tasks)? What treatments were effective for your symptoms? What treatments were ineffective?³

Social factors. Learn more about your patient’s support systems by asking: Who do you consider to be part of your social support? How is your relationship with your social support? Are there challenges in your relationship with your friends, family, or partner? If yes, what are those challenges? Are there other children in the home, and do you have support for them? Is your home environment safe? Do you feel that you have what you need for the baby? What stressors are you currently experiencing? Do you attend support groups for expectant mothers? Are you engaged in perinatal care?³

Continue to: Given the high prevalence...

Given the high prevalence of interpersonal violence in women of reproductive age, all patients should be screened for this. The American College of Obstetricians and Gynecologists Committee on Health Care for Underserved Women recommends screening for interpersonal violence at the first visit during the perinatal period, during each trimester, and at the postpartum visit (at minimum).⁴ Potential screening questions include (but are not limited to): Have you and/or your children ever been threatened by or felt afraid of your partner? When you argue with your partner, do either of you get physical? Has your partner ever physically hurt you (eg, hit, choked)? Do you feel safe at home? Do you have a safe place to go with resources you and your children will need in case of an emergency?^4-6

Feelings toward pregnancy, past/current pregnancy complications, and pregnancy loss. Ask your patient: Was this pregnancy planned? How do you feel about your pregnancy? How do you see yourself as a mother? Do you currently have pregnancy complications and/or have had them in the past, and, if so, what are/were they? Have you lost a pregnancy? If so, what was that like? Do you have fears related to childbirth, and, if so, what are they?³

Intrusive thoughts about harming the baby. Intrusive thoughts are common in postpartum women with anxiety disorders, including OCD.⁷ Merely asking patients if they’ve had thoughts of harming their baby is incomplete and insufficient to assess for intrusive thoughts. This question does not distinguish between intrusive thoughts and homicidal ideation; this distinction is absolutely necessary given the difference in potential risk to the infant.

Intrusive thoughts are generally associated with a low risk of mothers acting on their thoughts. These thoughts are typically ego dystonic and, in the most severe form, can be distressing to an extent that they cause behavioral changes, such as avoiding bathing the infant, avoiding diaper changes, avoiding knives, or separating themselves from the infant.⁷ On the contrary, having homicidal ideation carries a higher risk for harm to the infant. Homicidal ideation may be seen in patients with co-occurring psychosis, poor reality testing, and delusions.^5,7

Questions such as “Do you worry about harm coming to your baby?” “Do you worry about you causing harm to your baby?” and “Have you had an upsetting thought about harming your baby?” are more likely to reveal intrusive thoughts and prompt further exploration. Statements such as “Some people tell me that they have distressing thoughts about harm coming to their baby” can gently open the door to a having a dialogue about such thoughts. This dialogue is significantly important in making informed assessments as we develop comprehensive treatment plans.

Emerging data demonstrate that untreated perinatal anxiety is associated with negative outcomes, including an increased risk for suicide.¹ A 2017 systematic review and meta-analysis that included 102 studies with a total of 221,974 women from 34 countries found that the prevalence of self-reported anxiety symptoms and any anxiety disorder was 22.9% and 15.2%, respectively, across the 3 trimesters.¹ During pregnancy, anxiety disorders (eg, generalized anxiety disorder) and anxiety-related disorders (eg, obsessive-compulsive disorder [OCD] and posttraumatic stress disorder [PTSD]) can present as new illnesses or as a reoccurrence of an existing illness. Patients with pre-existing OCD may notice that the nature of their obsessions is changing. Women with pre-existing PTSD may have their symptoms triggered by pregnancy or delivery or may develop PTSD as a result of a traumatic delivery. Anxiety is frequently comorbid with depression, and high anxiety during pregnancy is one of the strongest risk factors for depression.^1,2

In light of this data, awareness and recognition of perinatal anxiety is critical. In this article, we describe how to accurately assess perinatal anxiety by avoiding assumptions and asking key questions during the clinical interview.

Avoid these common assumptions

Assessment begins with avoiding assumptions typically associated with maternal mental health. One common assumption is that pregnancy is a joyous occasion for all women. Pregnancy can be a stressful time that has its own unique difficulties, including the potential to develop or have a relapse of a mental illness. Another assumption is that the only concern is “postpartum depression.” In actuality, a significant percentage of women will experience depression during their pregnancy (not just in the postpartum period), and many other psychiatric illnesses are common during the perinatal period, including anxiety disorders.

Conduct a focused interview

Risk factors associated with antenatal anxiety include²:

• previous history of mental illness (particularly a history of anxiety and depression and a history of psychiatric treatment)
• lack of partner or social support
• history of abuse or domestic violence
• unplanned or unwanted pregnancy
• adverse events in life and high perceived stress
• present/past pregnancy complications
• pregnancy loss.

Symptoms of anxiety. The presence of anxiety or worrying does not necessarily mean a mother has an anxiety disorder. Using the DSM-5 as a guide, we should use the questions outlined in the following sections to inquire about all of the symptoms related to a particular illness, the pervasiveness of these symptoms, and to what extent these symptoms impair a woman’s ability to function and carry out her usual activities.³

Past psychiatric history. Ask your patient the following: Have you previously experienced anxiety and/or depressive symptoms? Were those symptoms limited only to times when you were pregnant or postpartum? Were your symptoms severe enough to disrupt your life (job, school, relationships, ability to complete daily tasks)? What treatments were effective for your symptoms? What treatments were ineffective?³

Social factors. Learn more about your patient’s support systems by asking: Who do you consider to be part of your social support? How is your relationship with your social support? Are there challenges in your relationship with your friends, family, or partner? If yes, what are those challenges? Are there other children in the home, and do you have support for them? Is your home environment safe? Do you feel that you have what you need for the baby? What stressors are you currently experiencing? Do you attend support groups for expectant mothers? Are you engaged in perinatal care?³

Continue to: Given the high prevalence...

Given the high prevalence of interpersonal violence in women of reproductive age, all patients should be screened for this. The American College of Obstetricians and Gynecologists Committee on Health Care for Underserved Women recommends screening for interpersonal violence at the first visit during the perinatal period, during each trimester, and at the postpartum visit (at minimum).⁴ Potential screening questions include (but are not limited to): Have you and/or your children ever been threatened by or felt afraid of your partner? When you argue with your partner, do either of you get physical? Has your partner ever physically hurt you (eg, hit, choked)? Do you feel safe at home? Do you have a safe place to go with resources you and your children will need in case of an emergency?^4-6

Feelings toward pregnancy, past/current pregnancy complications, and pregnancy loss. Ask your patient: Was this pregnancy planned? How do you feel about your pregnancy? How do you see yourself as a mother? Do you currently have pregnancy complications and/or have had them in the past, and, if so, what are/were they? Have you lost a pregnancy? If so, what was that like? Do you have fears related to childbirth, and, if so, what are they?³

Intrusive thoughts about harming the baby. Intrusive thoughts are common in postpartum women with anxiety disorders, including OCD.⁷ Merely asking patients if they’ve had thoughts of harming their baby is incomplete and insufficient to assess for intrusive thoughts. This question does not distinguish between intrusive thoughts and homicidal ideation; this distinction is absolutely necessary given the difference in potential risk to the infant.

Intrusive thoughts are generally associated with a low risk of mothers acting on their thoughts. These thoughts are typically ego dystonic and, in the most severe form, can be distressing to an extent that they cause behavioral changes, such as avoiding bathing the infant, avoiding diaper changes, avoiding knives, or separating themselves from the infant.⁷ On the contrary, having homicidal ideation carries a higher risk for harm to the infant. Homicidal ideation may be seen in patients with co-occurring psychosis, poor reality testing, and delusions.^5,7

Questions such as “Do you worry about harm coming to your baby?” “Do you worry about you causing harm to your baby?” and “Have you had an upsetting thought about harming your baby?” are more likely to reveal intrusive thoughts and prompt further exploration. Statements such as “Some people tell me that they have distressing thoughts about harm coming to their baby” can gently open the door to a having a dialogue about such thoughts. This dialogue is significantly important in making informed assessments as we develop comprehensive treatment plans.

Supportive psychotherapy is a common type of therapy that often is used in combination with other modalities. By focusing on improving symptoms and accepting the patient’s limitations, it is particularly helpful for individuals who might have difficulty engaging in insight-oriented psychotherapies, such as those struggling with external stressors, including exposure to trauma, bereavement, physical disabilities, or socioeconomic challenges. Personal limitations, including severe personality disorder or intellectual disabilities, might also limit a patient’s ability to self-reflect on subconscious issues, which can lead to choosing a supportive modality.

While being supportive in the vernacular sense can be helpful, formal supportive psychotherapy employs well-defined goals and techniques.¹ A therapist can facilitate progress by explicitly referring to these goals and techniques. The acronym ARISE can help therapists and other clinicians to use and appraise therapeutic progress toward these goals.

Alliance-building. The therapeutic alliance is an important predictor of the success of psychotherapy.² Warmly encourage positive transference toward the therapist. The patient’s appreciation of the therapist’s empathic interactions can further the alliance. Paraphrasing the patient’s words can demonstrate and enhance empathy. Doing so allows clarification of the patient’s thoughts and helps the patient feel understood. Formulate and partner around shared therapeutic goals. Monitor the strength of the alliance and intervene if it is threatened. For example, if you misunderstand your patient and inadvertently offend them, apologizing may be helpful. In the face of disagreement between the patient and therapist, reorienting back to shared goals reinforces common ground.

Reduce anxiety and negative affect. In contrast to the caricature of the stiff psychoanalyst, the supportive therapist adopts an engaged conversational style to help the patient feel relaxed and to diminish the power differential between therapist and patient. If the patient appears uncomfortable with silence, maintaining the flow of conversation may reduce discomfort.¹ Minimize your patient’s discomfort by approaching uncomfortable topics in manageable portions. Seek permission before introducing a subject that induces anxiety. Explain the reasoning behind approaching such topics.³ Reassurance and encouragement can further reduce anxiety.⁴ When not incongruous to the discussion, appropriate use of warm affect (eg, a smile) or even humor can elicit positive affect.

Increase awareness. Use psychoeducation and psychological interpretation (whether cognitive-behavioral or psychodynamic) to expand your patient’s awareness and help them understand their social contacts’ point of view. Clarification, gentle confrontation, and interpretation can make patients aware of biopsychosocial precipitants of distress.⁴

Strengthen coping mechanisms. Reinforce adaptive defense mechanisms, such as mature humor or suppression. Educating patients on practical organizational skills, problem-solving, relaxation techniques, and other relevant skills, can help them cope more effectively. Give advice only in limited circumstances, and when doing so, back up your advice with a rationale derived from your professional expertise. Because it is important for patients to realize that their life choices are their own, usually it is best to help the patient understand how they might come to their own decisions rather than to prescribe life choices in the form of advice.

Enhance self-esteem. Many patients in distress suffer from low self-esteem.^5,6 Active encouragement and honest praise can nurture your patient’s ability to correct a distorted self-image and challenge self-reproach. Praise should not be false but reality-based. Praise can address preexisting strengths, highlight the patient’s willingness to express challenging material, or provide reinforcement on progress made toward treatment goals.

Supportive psychotherapy is a common type of therapy that often is used in combination with other modalities. By focusing on improving symptoms and accepting the patient’s limitations, it is particularly helpful for individuals who might have difficulty engaging in insight-oriented psychotherapies, such as those struggling with external stressors, including exposure to trauma, bereavement, physical disabilities, or socioeconomic challenges. Personal limitations, including severe personality disorder or intellectual disabilities, might also limit a patient’s ability to self-reflect on subconscious issues, which can lead to choosing a supportive modality.

While being supportive in the vernacular sense can be helpful, formal supportive psychotherapy employs well-defined goals and techniques.¹ A therapist can facilitate progress by explicitly referring to these goals and techniques. The acronym ARISE can help therapists and other clinicians to use and appraise therapeutic progress toward these goals.

Alliance-building. The therapeutic alliance is an important predictor of the success of psychotherapy.² Warmly encourage positive transference toward the therapist. The patient’s appreciation of the therapist’s empathic interactions can further the alliance. Paraphrasing the patient’s words can demonstrate and enhance empathy. Doing so allows clarification of the patient’s thoughts and helps the patient feel understood. Formulate and partner around shared therapeutic goals. Monitor the strength of the alliance and intervene if it is threatened. For example, if you misunderstand your patient and inadvertently offend them, apologizing may be helpful. In the face of disagreement between the patient and therapist, reorienting back to shared goals reinforces common ground.

Reduce anxiety and negative affect. In contrast to the caricature of the stiff psychoanalyst, the supportive therapist adopts an engaged conversational style to help the patient feel relaxed and to diminish the power differential between therapist and patient. If the patient appears uncomfortable with silence, maintaining the flow of conversation may reduce discomfort.¹ Minimize your patient’s discomfort by approaching uncomfortable topics in manageable portions. Seek permission before introducing a subject that induces anxiety. Explain the reasoning behind approaching such topics.³ Reassurance and encouragement can further reduce anxiety.⁴ When not incongruous to the discussion, appropriate use of warm affect (eg, a smile) or even humor can elicit positive affect.

Increase awareness. Use psychoeducation and psychological interpretation (whether cognitive-behavioral or psychodynamic) to expand your patient’s awareness and help them understand their social contacts’ point of view. Clarification, gentle confrontation, and interpretation can make patients aware of biopsychosocial precipitants of distress.⁴

Strengthen coping mechanisms. Reinforce adaptive defense mechanisms, such as mature humor or suppression. Educating patients on practical organizational skills, problem-solving, relaxation techniques, and other relevant skills, can help them cope more effectively. Give advice only in limited circumstances, and when doing so, back up your advice with a rationale derived from your professional expertise. Because it is important for patients to realize that their life choices are their own, usually it is best to help the patient understand how they might come to their own decisions rather than to prescribe life choices in the form of advice.

Enhance self-esteem. Many patients in distress suffer from low self-esteem.^5,6 Active encouragement and honest praise can nurture your patient’s ability to correct a distorted self-image and challenge self-reproach. Praise should not be false but reality-based. Praise can address preexisting strengths, highlight the patient’s willingness to express challenging material, or provide reinforcement on progress made toward treatment goals.

Supportive psychotherapy is a common type of therapy that often is used in combination with other modalities. By focusing on improving symptoms and accepting the patient’s limitations, it is particularly helpful for individuals who might have difficulty engaging in insight-oriented psychotherapies, such as those struggling with external stressors, including exposure to trauma, bereavement, physical disabilities, or socioeconomic challenges. Personal limitations, including severe personality disorder or intellectual disabilities, might also limit a patient’s ability to self-reflect on subconscious issues, which can lead to choosing a supportive modality.

While being supportive in the vernacular sense can be helpful, formal supportive psychotherapy employs well-defined goals and techniques.¹ A therapist can facilitate progress by explicitly referring to these goals and techniques. The acronym ARISE can help therapists and other clinicians to use and appraise therapeutic progress toward these goals.

Alliance-building. The therapeutic alliance is an important predictor of the success of psychotherapy.² Warmly encourage positive transference toward the therapist. The patient’s appreciation of the therapist’s empathic interactions can further the alliance. Paraphrasing the patient’s words can demonstrate and enhance empathy. Doing so allows clarification of the patient’s thoughts and helps the patient feel understood. Formulate and partner around shared therapeutic goals. Monitor the strength of the alliance and intervene if it is threatened. For example, if you misunderstand your patient and inadvertently offend them, apologizing may be helpful. In the face of disagreement between the patient and therapist, reorienting back to shared goals reinforces common ground.

Reduce anxiety and negative affect. In contrast to the caricature of the stiff psychoanalyst, the supportive therapist adopts an engaged conversational style to help the patient feel relaxed and to diminish the power differential between therapist and patient. If the patient appears uncomfortable with silence, maintaining the flow of conversation may reduce discomfort.¹ Minimize your patient’s discomfort by approaching uncomfortable topics in manageable portions. Seek permission before introducing a subject that induces anxiety. Explain the reasoning behind approaching such topics.³ Reassurance and encouragement can further reduce anxiety.⁴ When not incongruous to the discussion, appropriate use of warm affect (eg, a smile) or even humor can elicit positive affect.

Increase awareness. Use psychoeducation and psychological interpretation (whether cognitive-behavioral or psychodynamic) to expand your patient’s awareness and help them understand their social contacts’ point of view. Clarification, gentle confrontation, and interpretation can make patients aware of biopsychosocial precipitants of distress.⁴

Strengthen coping mechanisms. Reinforce adaptive defense mechanisms, such as mature humor or suppression. Educating patients on practical organizational skills, problem-solving, relaxation techniques, and other relevant skills, can help them cope more effectively. Give advice only in limited circumstances, and when doing so, back up your advice with a rationale derived from your professional expertise. Because it is important for patients to realize that their life choices are their own, usually it is best to help the patient understand how they might come to their own decisions rather than to prescribe life choices in the form of advice.

Enhance self-esteem. Many patients in distress suffer from low self-esteem.^5,6 Active encouragement and honest praise can nurture your patient’s ability to correct a distorted self-image and challenge self-reproach. Praise should not be false but reality-based. Praise can address preexisting strengths, highlight the patient’s willingness to express challenging material, or provide reinforcement on progress made toward treatment goals.

DSM-5 defines personality disorders (PDs) as the presence of an enduring pattern of inner experience and behavior that “deviates markedly from the expectations of the individual’s culture, is pervasive and inflexible, has an onset in adulthood, is stable over time, and leads to distress or impairment.”¹ As a general rule, PDs are not limited to episodes of illness, but reflect an individual’s long-term adjustment. These disorders occur in 10% to 15% of the general population; the rates are especially high in health care settings, in criminal offenders, and in those with a substance use disorder (SUD).² PDs nearly always have an onset in adolescence or early adulthood and tend to diminish in severity with advancing age. They are associated with high rates of unemployment, homelessness, divorce and separation, domestic violence, substance misuse, and suicide.³

Psychotherapy is the first-line treatment for PDs, but there has been growing interest in using pharmacotherapy to treat PDs. While much of the PD treatment literature focuses on borderline PD,^4-9 this article describes diagnosis, potential pharmacotherapy strategies, and methods to assess response to treatment for patients with all types of PDs.

Recognizing and diagnosing personality disorders

The diagnosis of a PD requires an understanding of DSM-5 criteria combined with a comprehensive psychiatric history and mental status examination. The patient’s history is the most important basis for diagnosing a PD.² Collateral information from relatives or friends can help confirm the severity and pervasiveness of the individual’s personality problems. In some patients, long-term observation might be necessary to confirm the presence of a PD. Some clinicians are reluctant to diagnose PDs because of stigma, a problem common among patients with borderline PD.^10,11

To screen for PDs, a clinician might ask the patient about problems with interpersonal relationships, sense of self, work, affect, impulse control, and reality testing. Table 1¹² lists general screening questions for the presence of a PD from the Iowa Personality Disorders Screen. Structured diagnostic interviews and self-report assessments could boost recognition of PDs, but these tools are rarely used outside of research settings.^13,14

The PD clusters

DSM-5 divides 10 PDs into 3 clusters based on shared phenomenology and diagnostic criteria. Few patients have a “pure” case in which they meet criteria for only a single personality disorder.¹

Cluster A. “Eccentric cluster” disorders are united by social aversion, a failure to form close attachments, or paranoia and suspiciousness.¹⁵ These include paranoid, schizoid, and schizotypal PD. Low self-awareness is typical. There are no treatment guidelines for these disorders, although there is some clinical trial data for schizotypal PD.

Cluster B. “Dramatic cluster” disorders share dramatic, emotional, and erratic characteristics.¹⁴ These include narcissistic, antisocial, borderline, and histrionic PD. Antisocial and narcissistic patients have low self-awareness. There are treatment guidelines for antisocial and borderline PD, and a variety of clinical trial data is available for the latter.¹⁵

Continue to: Cluster C

Cluster C. “Anxious cluster” disorders are united by anxiousness, fearfulness, and poor self-esteem. Many of these patients also display interpersonal rigidity.¹⁵ These disorders include avoidant, dependent, and obsessive-compulsive PD. There are no treatment guidelines or clinical trial data for these disorders.

Why consider pharmacotherapy for personality disorders?

The consensus among experts is that psychotherapy is the treatment of choice for PDs.¹⁵ Despite significant gaps in the evidence base, there has been a growing interest in using psychotropic medication to treat PDs. For example, research shows that >90% of patients with borderline PD are prescribed medication, most typically antidepressants, antipsychotics, mood stabilizers, stimulants, or sedative-hypnotics.^16,17

Increased interest in pharmacotherapy for PDs could be related to research showing the importance of underlying neurobiology, particularly for antisocial and borderline PD.^18,19 This work is complemented by genetic research showing the heritability of PD traits and disorders.^20,21 Another factor could be renewed interest in dimensional approaches to the classification of PDs, as exemplified by DSM-5’s alternative model for PDs.¹ This approach aligns with some expert recommendations to focus on treating PD symptom dimensions, rather than the syndrome itself.²²

Importantly, no psychotropic medication is FDA-approved for the treatment of any PD. For that reason, prescribing medication for a PD is “off-label,” although prescribing a medication for a comorbid disorder for which the drug has an FDA-approved indication is not (eg, prescribing an antidepressant for major depressive disorder [MDD]).

Principles for prescribing

Despite gaps in research data, general principles for using medication to treat PDs have emerged from treatment guidelines for antisocial and borderline PD, clinical trial data, reviews and meta-analyses, and expert opinion. Clinicians should address the following considerations before prescribing medication to a patient with a PD.

Continue to: PD diagnosis

PD diagnosis. Has the patient been properly assessed and diagnosed? While history is the most important basis for diagnosis, the clinician should be familiar with the PDs and DSM-5 criteria. Has the patient been informed of the diagnosis and its implications for treatment?

Patient interest in medication. Is the patient interested in taking medication? Patients with borderline PD are often prescribed medication, but there are sparse data for the other PDs. The patient might have little interest in the PD diagnosis or its treatment.

Comorbidity. Has the patient been assessed for comorbid psychiatric disorders that could interfere with medication use (ie, an SUD) or might be a focus of treatment (eg, MDD)? Patients with PDs typically have significant comorbidity that a thorough evaluation will uncover.

PD symptom dimensions. Has the patient been assessed to determine cognitive or behavioral symptom dimensions of their PD? One or more symptom dimension(s) could be the focus of treatment. Table 2 lists examples of PD symptom dimensions.

Strategies to guide prescribing

Strategies to help guide prescribing include targeting any comorbid disorder(s), targeting important PD symptom dimensions (eg, impulsive aggression), choosing medication based on the similarity of the PD to another disorder known to respond to medication, and targeting the PD itself.

Continue to: Targeting comorbid disorders

Targeting comorbid disorders. National Institute for Health and Care Excellence guidelines for antisocial and borderline PD recommend that clinicians focus on treating comorbid disorders, a position echoed in Cochrane and other reviews.^4,9,22-26 For example, a patient with borderline PD experiencing a major depressive episode could be treated with an antidepressant. Targeting the depressive symptoms could boost the patient’s mood, perhaps lessening the individual’s PD symptoms or reducing their severity.

Targeting important symptom dimensions. For patients with borderline PD, several guidelines and reviews have suggested that treatment should focus on emotional dysregulation and impulsive aggression (mood stabilizers, antipsychotics), or cognitive-perceptual symptoms (antipsychotics).^4-6,15 There is some evidence that mood stabilizers or second-generation antipsychotics could help reduce impulsive aggression in patients with antisocial PD.²⁷

Choosing medication based on similarity to another disorder known to respond to medication. Avoidant PD overlaps with social anxiety disorder and can be conceptualized as a chronic, pervasive social phobia. Avoidant PD might respond to a medication known to be effective for treating social anxiety disorder, such as a selective serotonin reuptake inhibitor (SSRI) or venlafaxine.²⁸ Treating obsessive-compulsive PD with an SSRI is another example of this strategy, as 1 small study of fluvoxamine suggests.²⁹ Obsessive-compulsive PD is common in persons with obsessive-compulsive disorder, and overlap includes preoccupation with orders, rules, and lists, and an inability to throw things out.

Targeting the PD syndrome. Another strategy is to target the PD itself. Clinical trial data suggest the antipsychotic risperidone can reduce the symptoms of schizotypal PD.³⁰ Considering that this PD has a genetic association with schizophrenia, it is not surprising that the patient’s ideas of reference, odd communication, or transient paranoia might respond to an antipsychotic. Data from randomized controlled trials (RCTs) support the use of the second-generation antipsychotics aripiprazole and quetiapine to treat BPD.^31,32 While older guidelines^4,5 supported the use of the mood stabilizer lamotrigine, a recent RCT found that it was no more effective than placebo for borderline PD or its symptom dimensions.³³

What to do before prescribing

Before writing a prescription, the clinician and patient should discuss the presence of a PD and the desirability of treatment. The patient should understand the limited evidence base and know that medication prescribed for a PD is off-label. The clinician should discuss medication selection and its rationale, and whether the medication is targeting a comorbid disorder, symptom dimension(s), or the PD itself. Additional considerations for prescribing for patients with PDs are listed in Table 3.³⁴

Continue to: Avoid polypharmacy

Avoid polypharmacy. Many patients with borderline PD are prescribed multiple psychotropic medications.^16,17 This approach leads to greater expense and more adverse effects, and is not evidence-based.

Avoid benzodiazepines. Many patients with borderline PD are prescribed benzodiazepines, often as part of a polypharmacy regimen. These drugs can cause disinhibition, thereby increasing acting-out behaviors and self-harm.³⁵ Also, patients with PDs often have SUDs, which is a contraindication for benzodiazepine use.

Rate the patient’s improvement. Both the patient and clinician can benefit from monitoring symptomatic improvement. Several validated scales can be used to rate depression, anxiety, impulsivity, mood lability, anger, and aggression (Table 4^36-41).Some validated scales for borderline PD align with DSM-5 criteria. Two such widely used instruments are the Zanarini Rating Scale for Borderline Personality Disorder (ZAN-BPD)⁴² and the self-rated Borderline Evaluation of Severity Over Time (BEST).⁴³ Each has questions that could be pulled to rate a symptom dimension of interest, such as affective instability, anger dyscontrol, or abandonment fears (Table 5^42,43).

A visual analog scale is easy to use and can target symptom dimensions of interest.⁴⁴ For example, a clinician could use a visual analog scale to rate mood instability by asking a patient to rate their mood severity by making a mark along a 10-cm line (0 = “Most erratic emotions I have experienced,” 10 = “Most stable I have ever experienced my emotions to be”). This score can be recorded at baseline and subsequent visits.

Take-home points

PDs are common in the general population and health care settings. They are underrecognized by the general public and mental health professionals, often because of stigma. Clinicians could boost their recognition of these disorders by embedding simple screening questions in their patient assessments. Many patients with PDs will be interested in pharmacotherapy for their disorder or symptoms. Treatment strategies include targeting the comorbid disorder(s), targeting important PD symptom dimensions, choosing medication based on the similarity of the PD to another disorder known to respond to medication, and targeting the PD itself. Each strategy has its limitations and varying degrees of empirical support. Treatment response can be monitored using validated scales or a visual analog scale.

Continue to: Bottom Line

Bottom Line

Although psychotherapy is the first-line treatment and no medications are FDAapproved for treating personality disorders (PDs), there has been growing interest in using psychotropic medication to treat PDs. Strategies for pharmacotherapy include targeting comorbid disorders, PD symptom dimensions, or the PD itself. Choice of medication can be based on the similarity of the PD with another disorder known to respond to medication.

Related Resources

• Correa Da Costa S, Sanches M, Soares JC. Bipolar disorder or borderline personality disorder? Current Psychiatry. 2019;18(11):26-29,35-39.
• Bateman A, Gunderson J, Mulder R. Treatment of personality disorders. Lancet. 2015;385(9969):735-743.

Drug Brand Names

Aripiprazole • Abilify
Fluvoxamine • Luvox
Lamotrigine • Lamictal
Quetiapine • Seroquel
Risperidone • Risperdal
Venlafaxine • Effexor

DSM-5 defines personality disorders (PDs) as the presence of an enduring pattern of inner experience and behavior that “deviates markedly from the expectations of the individual’s culture, is pervasive and inflexible, has an onset in adulthood, is stable over time, and leads to distress or impairment.”¹ As a general rule, PDs are not limited to episodes of illness, but reflect an individual’s long-term adjustment. These disorders occur in 10% to 15% of the general population; the rates are especially high in health care settings, in criminal offenders, and in those with a substance use disorder (SUD).² PDs nearly always have an onset in adolescence or early adulthood and tend to diminish in severity with advancing age. They are associated with high rates of unemployment, homelessness, divorce and separation, domestic violence, substance misuse, and suicide.³

Psychotherapy is the first-line treatment for PDs, but there has been growing interest in using pharmacotherapy to treat PDs. While much of the PD treatment literature focuses on borderline PD,^4-9 this article describes diagnosis, potential pharmacotherapy strategies, and methods to assess response to treatment for patients with all types of PDs.

Recognizing and diagnosing personality disorders

The diagnosis of a PD requires an understanding of DSM-5 criteria combined with a comprehensive psychiatric history and mental status examination. The patient’s history is the most important basis for diagnosing a PD.² Collateral information from relatives or friends can help confirm the severity and pervasiveness of the individual’s personality problems. In some patients, long-term observation might be necessary to confirm the presence of a PD. Some clinicians are reluctant to diagnose PDs because of stigma, a problem common among patients with borderline PD.^10,11

To screen for PDs, a clinician might ask the patient about problems with interpersonal relationships, sense of self, work, affect, impulse control, and reality testing. Table 1¹² lists general screening questions for the presence of a PD from the Iowa Personality Disorders Screen. Structured diagnostic interviews and self-report assessments could boost recognition of PDs, but these tools are rarely used outside of research settings.^13,14

The PD clusters

DSM-5 divides 10 PDs into 3 clusters based on shared phenomenology and diagnostic criteria. Few patients have a “pure” case in which they meet criteria for only a single personality disorder.¹

Cluster A. “Eccentric cluster” disorders are united by social aversion, a failure to form close attachments, or paranoia and suspiciousness.¹⁵ These include paranoid, schizoid, and schizotypal PD. Low self-awareness is typical. There are no treatment guidelines for these disorders, although there is some clinical trial data for schizotypal PD.

Cluster B. “Dramatic cluster” disorders share dramatic, emotional, and erratic characteristics.¹⁴ These include narcissistic, antisocial, borderline, and histrionic PD. Antisocial and narcissistic patients have low self-awareness. There are treatment guidelines for antisocial and borderline PD, and a variety of clinical trial data is available for the latter.¹⁵

Continue to: Cluster C

Cluster C. “Anxious cluster” disorders are united by anxiousness, fearfulness, and poor self-esteem. Many of these patients also display interpersonal rigidity.¹⁵ These disorders include avoidant, dependent, and obsessive-compulsive PD. There are no treatment guidelines or clinical trial data for these disorders.

Why consider pharmacotherapy for personality disorders?

The consensus among experts is that psychotherapy is the treatment of choice for PDs.¹⁵ Despite significant gaps in the evidence base, there has been a growing interest in using psychotropic medication to treat PDs. For example, research shows that >90% of patients with borderline PD are prescribed medication, most typically antidepressants, antipsychotics, mood stabilizers, stimulants, or sedative-hypnotics.^16,17

Increased interest in pharmacotherapy for PDs could be related to research showing the importance of underlying neurobiology, particularly for antisocial and borderline PD.^18,19 This work is complemented by genetic research showing the heritability of PD traits and disorders.^20,21 Another factor could be renewed interest in dimensional approaches to the classification of PDs, as exemplified by DSM-5’s alternative model for PDs.¹ This approach aligns with some expert recommendations to focus on treating PD symptom dimensions, rather than the syndrome itself.²²

Importantly, no psychotropic medication is FDA-approved for the treatment of any PD. For that reason, prescribing medication for a PD is “off-label,” although prescribing a medication for a comorbid disorder for which the drug has an FDA-approved indication is not (eg, prescribing an antidepressant for major depressive disorder [MDD]).

Principles for prescribing

Despite gaps in research data, general principles for using medication to treat PDs have emerged from treatment guidelines for antisocial and borderline PD, clinical trial data, reviews and meta-analyses, and expert opinion. Clinicians should address the following considerations before prescribing medication to a patient with a PD.

Continue to: PD diagnosis

PD diagnosis. Has the patient been properly assessed and diagnosed? While history is the most important basis for diagnosis, the clinician should be familiar with the PDs and DSM-5 criteria. Has the patient been informed of the diagnosis and its implications for treatment?

Patient interest in medication. Is the patient interested in taking medication? Patients with borderline PD are often prescribed medication, but there are sparse data for the other PDs. The patient might have little interest in the PD diagnosis or its treatment.

Comorbidity. Has the patient been assessed for comorbid psychiatric disorders that could interfere with medication use (ie, an SUD) or might be a focus of treatment (eg, MDD)? Patients with PDs typically have significant comorbidity that a thorough evaluation will uncover.

PD symptom dimensions. Has the patient been assessed to determine cognitive or behavioral symptom dimensions of their PD? One or more symptom dimension(s) could be the focus of treatment. Table 2 lists examples of PD symptom dimensions.

Strategies to guide prescribing

Strategies to help guide prescribing include targeting any comorbid disorder(s), targeting important PD symptom dimensions (eg, impulsive aggression), choosing medication based on the similarity of the PD to another disorder known to respond to medication, and targeting the PD itself.

Continue to: Targeting comorbid disorders

Targeting comorbid disorders. National Institute for Health and Care Excellence guidelines for antisocial and borderline PD recommend that clinicians focus on treating comorbid disorders, a position echoed in Cochrane and other reviews.^4,9,22-26 For example, a patient with borderline PD experiencing a major depressive episode could be treated with an antidepressant. Targeting the depressive symptoms could boost the patient’s mood, perhaps lessening the individual’s PD symptoms or reducing their severity.

Targeting important symptom dimensions. For patients with borderline PD, several guidelines and reviews have suggested that treatment should focus on emotional dysregulation and impulsive aggression (mood stabilizers, antipsychotics), or cognitive-perceptual symptoms (antipsychotics).^4-6,15 There is some evidence that mood stabilizers or second-generation antipsychotics could help reduce impulsive aggression in patients with antisocial PD.²⁷

Choosing medication based on similarity to another disorder known to respond to medication. Avoidant PD overlaps with social anxiety disorder and can be conceptualized as a chronic, pervasive social phobia. Avoidant PD might respond to a medication known to be effective for treating social anxiety disorder, such as a selective serotonin reuptake inhibitor (SSRI) or venlafaxine.²⁸ Treating obsessive-compulsive PD with an SSRI is another example of this strategy, as 1 small study of fluvoxamine suggests.²⁹ Obsessive-compulsive PD is common in persons with obsessive-compulsive disorder, and overlap includes preoccupation with orders, rules, and lists, and an inability to throw things out.

Targeting the PD syndrome. Another strategy is to target the PD itself. Clinical trial data suggest the antipsychotic risperidone can reduce the symptoms of schizotypal PD.³⁰ Considering that this PD has a genetic association with schizophrenia, it is not surprising that the patient’s ideas of reference, odd communication, or transient paranoia might respond to an antipsychotic. Data from randomized controlled trials (RCTs) support the use of the second-generation antipsychotics aripiprazole and quetiapine to treat BPD.^31,32 While older guidelines^4,5 supported the use of the mood stabilizer lamotrigine, a recent RCT found that it was no more effective than placebo for borderline PD or its symptom dimensions.³³

What to do before prescribing

Before writing a prescription, the clinician and patient should discuss the presence of a PD and the desirability of treatment. The patient should understand the limited evidence base and know that medication prescribed for a PD is off-label. The clinician should discuss medication selection and its rationale, and whether the medication is targeting a comorbid disorder, symptom dimension(s), or the PD itself. Additional considerations for prescribing for patients with PDs are listed in Table 3.³⁴

Continue to: Avoid polypharmacy

Avoid polypharmacy. Many patients with borderline PD are prescribed multiple psychotropic medications.^16,17 This approach leads to greater expense and more adverse effects, and is not evidence-based.

Avoid benzodiazepines. Many patients with borderline PD are prescribed benzodiazepines, often as part of a polypharmacy regimen. These drugs can cause disinhibition, thereby increasing acting-out behaviors and self-harm.³⁵ Also, patients with PDs often have SUDs, which is a contraindication for benzodiazepine use.

Rate the patient’s improvement. Both the patient and clinician can benefit from monitoring symptomatic improvement. Several validated scales can be used to rate depression, anxiety, impulsivity, mood lability, anger, and aggression (Table 4^36-41).Some validated scales for borderline PD align with DSM-5 criteria. Two such widely used instruments are the Zanarini Rating Scale for Borderline Personality Disorder (ZAN-BPD)⁴² and the self-rated Borderline Evaluation of Severity Over Time (BEST).⁴³ Each has questions that could be pulled to rate a symptom dimension of interest, such as affective instability, anger dyscontrol, or abandonment fears (Table 5^42,43).

A visual analog scale is easy to use and can target symptom dimensions of interest.⁴⁴ For example, a clinician could use a visual analog scale to rate mood instability by asking a patient to rate their mood severity by making a mark along a 10-cm line (0 = “Most erratic emotions I have experienced,” 10 = “Most stable I have ever experienced my emotions to be”). This score can be recorded at baseline and subsequent visits.

Take-home points

PDs are common in the general population and health care settings. They are underrecognized by the general public and mental health professionals, often because of stigma. Clinicians could boost their recognition of these disorders by embedding simple screening questions in their patient assessments. Many patients with PDs will be interested in pharmacotherapy for their disorder or symptoms. Treatment strategies include targeting the comorbid disorder(s), targeting important PD symptom dimensions, choosing medication based on the similarity of the PD to another disorder known to respond to medication, and targeting the PD itself. Each strategy has its limitations and varying degrees of empirical support. Treatment response can be monitored using validated scales or a visual analog scale.

Continue to: Bottom Line

Bottom Line

Although psychotherapy is the first-line treatment and no medications are FDAapproved for treating personality disorders (PDs), there has been growing interest in using psychotropic medication to treat PDs. Strategies for pharmacotherapy include targeting comorbid disorders, PD symptom dimensions, or the PD itself. Choice of medication can be based on the similarity of the PD with another disorder known to respond to medication.

Related Resources

• Correa Da Costa S, Sanches M, Soares JC. Bipolar disorder or borderline personality disorder? Current Psychiatry. 2019;18(11):26-29,35-39.
• Bateman A, Gunderson J, Mulder R. Treatment of personality disorders. Lancet. 2015;385(9969):735-743.

Drug Brand Names

Aripiprazole • Abilify
Fluvoxamine • Luvox
Lamotrigine • Lamictal
Quetiapine • Seroquel
Risperidone • Risperdal
Venlafaxine • Effexor

DSM-5 defines personality disorders (PDs) as the presence of an enduring pattern of inner experience and behavior that “deviates markedly from the expectations of the individual’s culture, is pervasive and inflexible, has an onset in adulthood, is stable over time, and leads to distress or impairment.”¹ As a general rule, PDs are not limited to episodes of illness, but reflect an individual’s long-term adjustment. These disorders occur in 10% to 15% of the general population; the rates are especially high in health care settings, in criminal offenders, and in those with a substance use disorder (SUD).² PDs nearly always have an onset in adolescence or early adulthood and tend to diminish in severity with advancing age. They are associated with high rates of unemployment, homelessness, divorce and separation, domestic violence, substance misuse, and suicide.³

Psychotherapy is the first-line treatment for PDs, but there has been growing interest in using pharmacotherapy to treat PDs. While much of the PD treatment literature focuses on borderline PD,^4-9 this article describes diagnosis, potential pharmacotherapy strategies, and methods to assess response to treatment for patients with all types of PDs.

Recognizing and diagnosing personality disorders

The diagnosis of a PD requires an understanding of DSM-5 criteria combined with a comprehensive psychiatric history and mental status examination. The patient’s history is the most important basis for diagnosing a PD.² Collateral information from relatives or friends can help confirm the severity and pervasiveness of the individual’s personality problems. In some patients, long-term observation might be necessary to confirm the presence of a PD. Some clinicians are reluctant to diagnose PDs because of stigma, a problem common among patients with borderline PD.^10,11

To screen for PDs, a clinician might ask the patient about problems with interpersonal relationships, sense of self, work, affect, impulse control, and reality testing. Table 1¹² lists general screening questions for the presence of a PD from the Iowa Personality Disorders Screen. Structured diagnostic interviews and self-report assessments could boost recognition of PDs, but these tools are rarely used outside of research settings.^13,14

The PD clusters

DSM-5 divides 10 PDs into 3 clusters based on shared phenomenology and diagnostic criteria. Few patients have a “pure” case in which they meet criteria for only a single personality disorder.¹

Cluster A. “Eccentric cluster” disorders are united by social aversion, a failure to form close attachments, or paranoia and suspiciousness.¹⁵ These include paranoid, schizoid, and schizotypal PD. Low self-awareness is typical. There are no treatment guidelines for these disorders, although there is some clinical trial data for schizotypal PD.

Cluster B. “Dramatic cluster” disorders share dramatic, emotional, and erratic characteristics.¹⁴ These include narcissistic, antisocial, borderline, and histrionic PD. Antisocial and narcissistic patients have low self-awareness. There are treatment guidelines for antisocial and borderline PD, and a variety of clinical trial data is available for the latter.¹⁵

Continue to: Cluster C

Cluster C. “Anxious cluster” disorders are united by anxiousness, fearfulness, and poor self-esteem. Many of these patients also display interpersonal rigidity.¹⁵ These disorders include avoidant, dependent, and obsessive-compulsive PD. There are no treatment guidelines or clinical trial data for these disorders.

Why consider pharmacotherapy for personality disorders?

The consensus among experts is that psychotherapy is the treatment of choice for PDs.¹⁵ Despite significant gaps in the evidence base, there has been a growing interest in using psychotropic medication to treat PDs. For example, research shows that >90% of patients with borderline PD are prescribed medication, most typically antidepressants, antipsychotics, mood stabilizers, stimulants, or sedative-hypnotics.^16,17

Increased interest in pharmacotherapy for PDs could be related to research showing the importance of underlying neurobiology, particularly for antisocial and borderline PD.^18,19 This work is complemented by genetic research showing the heritability of PD traits and disorders.^20,21 Another factor could be renewed interest in dimensional approaches to the classification of PDs, as exemplified by DSM-5’s alternative model for PDs.¹ This approach aligns with some expert recommendations to focus on treating PD symptom dimensions, rather than the syndrome itself.²²

Importantly, no psychotropic medication is FDA-approved for the treatment of any PD. For that reason, prescribing medication for a PD is “off-label,” although prescribing a medication for a comorbid disorder for which the drug has an FDA-approved indication is not (eg, prescribing an antidepressant for major depressive disorder [MDD]).

Principles for prescribing

Despite gaps in research data, general principles for using medication to treat PDs have emerged from treatment guidelines for antisocial and borderline PD, clinical trial data, reviews and meta-analyses, and expert opinion. Clinicians should address the following considerations before prescribing medication to a patient with a PD.

Continue to: PD diagnosis

PD diagnosis. Has the patient been properly assessed and diagnosed? While history is the most important basis for diagnosis, the clinician should be familiar with the PDs and DSM-5 criteria. Has the patient been informed of the diagnosis and its implications for treatment?

Patient interest in medication. Is the patient interested in taking medication? Patients with borderline PD are often prescribed medication, but there are sparse data for the other PDs. The patient might have little interest in the PD diagnosis or its treatment.

Comorbidity. Has the patient been assessed for comorbid psychiatric disorders that could interfere with medication use (ie, an SUD) or might be a focus of treatment (eg, MDD)? Patients with PDs typically have significant comorbidity that a thorough evaluation will uncover.

PD symptom dimensions. Has the patient been assessed to determine cognitive or behavioral symptom dimensions of their PD? One or more symptom dimension(s) could be the focus of treatment. Table 2 lists examples of PD symptom dimensions.

Strategies to guide prescribing

Strategies to help guide prescribing include targeting any comorbid disorder(s), targeting important PD symptom dimensions (eg, impulsive aggression), choosing medication based on the similarity of the PD to another disorder known to respond to medication, and targeting the PD itself.

Continue to: Targeting comorbid disorders

Targeting comorbid disorders. National Institute for Health and Care Excellence guidelines for antisocial and borderline PD recommend that clinicians focus on treating comorbid disorders, a position echoed in Cochrane and other reviews.^4,9,22-26 For example, a patient with borderline PD experiencing a major depressive episode could be treated with an antidepressant. Targeting the depressive symptoms could boost the patient’s mood, perhaps lessening the individual’s PD symptoms or reducing their severity.

Targeting important symptom dimensions. For patients with borderline PD, several guidelines and reviews have suggested that treatment should focus on emotional dysregulation and impulsive aggression (mood stabilizers, antipsychotics), or cognitive-perceptual symptoms (antipsychotics).^4-6,15 There is some evidence that mood stabilizers or second-generation antipsychotics could help reduce impulsive aggression in patients with antisocial PD.²⁷

Choosing medication based on similarity to another disorder known to respond to medication. Avoidant PD overlaps with social anxiety disorder and can be conceptualized as a chronic, pervasive social phobia. Avoidant PD might respond to a medication known to be effective for treating social anxiety disorder, such as a selective serotonin reuptake inhibitor (SSRI) or venlafaxine.²⁸ Treating obsessive-compulsive PD with an SSRI is another example of this strategy, as 1 small study of fluvoxamine suggests.²⁹ Obsessive-compulsive PD is common in persons with obsessive-compulsive disorder, and overlap includes preoccupation with orders, rules, and lists, and an inability to throw things out.

Targeting the PD syndrome. Another strategy is to target the PD itself. Clinical trial data suggest the antipsychotic risperidone can reduce the symptoms of schizotypal PD.³⁰ Considering that this PD has a genetic association with schizophrenia, it is not surprising that the patient’s ideas of reference, odd communication, or transient paranoia might respond to an antipsychotic. Data from randomized controlled trials (RCTs) support the use of the second-generation antipsychotics aripiprazole and quetiapine to treat BPD.^31,32 While older guidelines^4,5 supported the use of the mood stabilizer lamotrigine, a recent RCT found that it was no more effective than placebo for borderline PD or its symptom dimensions.³³

What to do before prescribing

Before writing a prescription, the clinician and patient should discuss the presence of a PD and the desirability of treatment. The patient should understand the limited evidence base and know that medication prescribed for a PD is off-label. The clinician should discuss medication selection and its rationale, and whether the medication is targeting a comorbid disorder, symptom dimension(s), or the PD itself. Additional considerations for prescribing for patients with PDs are listed in Table 3.³⁴

Continue to: Avoid polypharmacy

Avoid polypharmacy. Many patients with borderline PD are prescribed multiple psychotropic medications.^16,17 This approach leads to greater expense and more adverse effects, and is not evidence-based.

Avoid benzodiazepines. Many patients with borderline PD are prescribed benzodiazepines, often as part of a polypharmacy regimen. These drugs can cause disinhibition, thereby increasing acting-out behaviors and self-harm.³⁵ Also, patients with PDs often have SUDs, which is a contraindication for benzodiazepine use.

Rate the patient’s improvement. Both the patient and clinician can benefit from monitoring symptomatic improvement. Several validated scales can be used to rate depression, anxiety, impulsivity, mood lability, anger, and aggression (Table 4^36-41).Some validated scales for borderline PD align with DSM-5 criteria. Two such widely used instruments are the Zanarini Rating Scale for Borderline Personality Disorder (ZAN-BPD)⁴² and the self-rated Borderline Evaluation of Severity Over Time (BEST).⁴³ Each has questions that could be pulled to rate a symptom dimension of interest, such as affective instability, anger dyscontrol, or abandonment fears (Table 5^42,43).

A visual analog scale is easy to use and can target symptom dimensions of interest.⁴⁴ For example, a clinician could use a visual analog scale to rate mood instability by asking a patient to rate their mood severity by making a mark along a 10-cm line (0 = “Most erratic emotions I have experienced,” 10 = “Most stable I have ever experienced my emotions to be”). This score can be recorded at baseline and subsequent visits.

Take-home points

PDs are common in the general population and health care settings. They are underrecognized by the general public and mental health professionals, often because of stigma. Clinicians could boost their recognition of these disorders by embedding simple screening questions in their patient assessments. Many patients with PDs will be interested in pharmacotherapy for their disorder or symptoms. Treatment strategies include targeting the comorbid disorder(s), targeting important PD symptom dimensions, choosing medication based on the similarity of the PD to another disorder known to respond to medication, and targeting the PD itself. Each strategy has its limitations and varying degrees of empirical support. Treatment response can be monitored using validated scales or a visual analog scale.

Continue to: Bottom Line

Bottom Line

Although psychotherapy is the first-line treatment and no medications are FDAapproved for treating personality disorders (PDs), there has been growing interest in using psychotropic medication to treat PDs. Strategies for pharmacotherapy include targeting comorbid disorders, PD symptom dimensions, or the PD itself. Choice of medication can be based on the similarity of the PD with another disorder known to respond to medication.

Related Resources

• Correa Da Costa S, Sanches M, Soares JC. Bipolar disorder or borderline personality disorder? Current Psychiatry. 2019;18(11):26-29,35-39.
• Bateman A, Gunderson J, Mulder R. Treatment of personality disorders. Lancet. 2015;385(9969):735-743.

Drug Brand Names

Aripiprazole • Abilify
Fluvoxamine • Luvox
Lamotrigine • Lamictal
Quetiapine • Seroquel
Risperidone • Risperdal
Venlafaxine • Effexor

Ms. N, age 30, presents to the emergency department for altered mental status, insomnia, and behavioral changes, which she has experienced for 1 week. On evaluation, she grabs a clinician’s hand and details her business ideas and life story with no prompting. Ms. N’s mental status examination is significant for hyperverbal speech with increased rate and volume; tangential thought process; and bright, expanded affect.

One week earlier, Ms. N was hospitalized for sudden-onset chest pain, weakness, and dizziness. She received 45 minutes of cardiopulmonary resuscitation prior to presentation and was found to have a ST-segment elevation myocardial infarction that required emergent left anterior descending coronary artery and right coronary artery percutaneous coronary intervention to place drug-eluting stents. Her recovery was complicated by acute cardiogenic shock, pulmonary edema, and hypoxic respiratory failure. Subsequently, she was intubated, admitted to the ICU, and received high-dose corticosteroids, including IV methylprednisolone, 40 mg every 12 hours, which was tapered prior to discharge. Her husband reports that since Ms. N came home, she has been more talkative and irritable, ruminating about past events, unable to sleep (<1 hour/night), and crying frequently. She has also been endorsing visual and auditory hallucinations, with increased praying and listening to religious music.

The frequent clinical use of steroids necessitates an understanding of these medications’ various adverse effects. The manifestations of steroid-induced psychiatric symptoms are broad and can involve affective, behavioral, and cognitive domains. While the current mechanism is unknown, this phenomenon may be related to decreased levels of corticotropin, norepinephrine, and beta-endorphin immunoreactivity, as well as effects on brain regions such as the hippocampus and amygdala. The best interventions for steroid-induced psychiatric symptoms are awareness and early diagnosis. There are no FDA-approved treatments for steroid-induced psychiatric symptoms; initial measures should include tapering or discontinuing corticosteroids.

In this article, we review the literature on the incidence, characteristics, differential diagnoses, proposed mechanism, risk factors, and proposed treatments of steroid-induced psychiatric symptoms.

A wide range of presentations

Steroid use has increased over the past 2 decades, with 10% of medical and surgical inpatients and 1% to 3% of the general population taking long-term glucocorticoids.¹ Even with topical application, steroid therapy is often systemically absorbed, and thus may lead to steroid-induced psychiatric symptoms. The incidence of steroid-induced psychiatric symptoms is difficult to assess because there can be a wide range of reactions that are dose- and time-related. Three reviews of a total of 122 cases reports found that an estimated 5% of patients treated with steroids experience severe psychiatric reactions.^1-3

Steroid-induced psychopathology can include mood, behavioral, and/or cognitive impairments. Mania/hypomania is the most common overall psychiatric symptom; the most common mood manifestations are anxiety and depression.^4,5 Other possible steroid-induced symptoms include psychosis, dementia, panic disorder, delirium, suicidal thinking and behavior, aggressive behavior, insomnia, agitation, depersonalization, and euphoria.⁵ The most common cognitive impairment is verbal or declarative memory deficit; others include distractibility and deficits in attention and psychomotor speed.⁵ These psychiatric symptoms can have a rapid onset, possibly within hours of starting steroids.¹ However, studies have reported a median time to onset of 11.5 days; 39% of cases had onset during the first week and 62% within 2 weeks.^3,6 After reducing or stopping the steroid, it may take days to weeks before symptoms start to subside.²

What to consider in the differential Dx

Psychiatric symptoms that are induced by steroids can mimic metabolic, neurologic, or toxic disorders. Other factors to consider include drug withdrawal/intoxication, infections, and paraneoplastic syndromes.^4,5 Although there is no reported correlation between the location of neurologic lesions and the development of specific psychiatric symptoms, manic symptoms appear most commonly with lesions in the right frontal lobe. ⁴ Other factors to note include the presence of new-onset psychiatric illnesses such as bipolar, mood, or thought disorders,⁴ as well as psychosocial stressors that might be contributing to the patient’s presentation.⁵

Continue to: Proposed mechanisms

Although the exact mechanism by which steroids induce psychiatric symptoms is unknown, several mechanisms have been proposed. One hypothesis is that steroid-induced psychopathology is related to decreased levels of corticotropin, norepinephrine, and beta-endorphin immunoreactivity.^4,5,7 This may explain why many patients with major depressive disorder have elevated cortisol production and/or lack of suppression of cortisol secretion during a dexamethasone stimulation test, and why approximately one-half of patients with Cushing’s disease experience depressive symptoms.⁸ This is also likely why antipsychotics, which typically reduce cortisol, are efficacious treatments for some steroid-induced psychiatric symptoms.⁹ 

Cognitive impairments from steroid use may be related to these agents’ effects on certain brain regions. One such area is the hippocampus, an important mediator in the creation and maintenance of episodic and declarative memories.^5,8,9 Acute glucocorticoid use is associated with decreased activity in the left hippocampus, reduced hippocampal glucose metabolism, and reduced cerebral blood flow in the posterior medial temporal lobe.¹⁰ Long-term glucocorticoid exposure is associated with smaller hippocampal volume and lower levels of temporal lobe N-acetylaspartate, a marker of neuronal viability.¹⁰ Because working memory depends on the prefrontal cortex and declarative memory relies on the hippocampus, deficits in these functions can be attributed to the effect of prolonged glucocorticoid exposure on glucocorticoid or mineralocorticoid receptors in the hippocampus, reduction of hippocampal volume, or elevated glutamate accumulation in that area.¹¹ In addition, high cortisol levels inhibit brain-derived neurotrophic factor, which plays a crucial role in maintaining neural architecture in key brain regions such as the hippocampus and prefrontal cortex.¹¹ There is also a correlation between the duration of prednisone treatment and atrophy of the right amygdala, which is an important regulator of mood and anxiety.¹¹ Both the hippocampus and amygdala have dense collections of glucocorticoid receptors. This may explain why patients who receive high-dose corticosteroids can have reversible atrophy in the hypothalamus and amygdala, leading to deficits in emotional learning and the stress response.

Factors that increase risk

Several factors can increase the risk of steroid-induced psychopathology. The most significant is the dose; higher doses are more likely to produce psychiatric symptoms.^1,5 Concurrent use of drugs that increase circulating levels of corticosteroids, such as inhibitors of the cytochrome P450 (CYP) enzyme (eg, clarithromycin), also increases the likelihood of developing psychiatric symptoms.^1,5 Risk is also increased in patients with liver or renal dysfunction.¹ Cerebral spinal fluid/serum albumin ratio, a marker of blood-brain barrier damage, and low serum complement levels were also reported to be independent risk factors,¹² with the thought that increased permeability of the blood-brain barrier may allow hydrophobic steroid molecules to more easily penetrate the CNS, leading to increased neuropsychiatric effects. Hypoalbuminemia is another reported risk factor, perhaps because lower levels of serum albumin are related to higher levels of free and active glucocorticoids, which are normally inactive when bound to albumin.¹³ There also appears to be an increased prevalence of steroid-induced psychopathology in women, perhaps due to greater propensity in women to seek medical care or a higher prevalence of women with medical disorders that are treated with steroids.⁵ A previous history of psychiatric disorders may not increase risk.⁵

Several methods for reducing risk have been proposed, including using a divided-dosing regimens that may lower peak steroid plasma concentrations.^13,14 However, the best prevention of steroid-induced psychiatric symptoms are awareness, early diagnosis, and intervention. Studies have suggested that N-methyl-d-aspartate (NMDA) antagonists¹⁵ and other agents that decrease glutamate release (such as phenytoin and lamotrigine¹⁶) may help prevent corticosteroid-induced hippocampal volume loss. Lamotrigine has been shown to reduce the amount of atrophy in the amygdala in patients taking corticosteroids.¹⁷ Phenytoin has also been reported to reduce the incidence of hypomania associated with corticosteroids, perhaps due to its induction of CYP450 activity and acceleration of steroid clearance.¹⁶

Treatment options

There are no FDA-approved medications for managing steroid-induced psychiatric symptoms.^1,16 Treatment is based on evidence from case reports and a few small case series (Table^2-5,17,18).

Continue to: When possible, initial treatment...

When possible, initial treatment should include discontinuing or tapering corticosteroids to <40 mg/d of prednisone-equivalent.^1,4,10,18 Most studies have reported rapid reversal of deficits in declarative memory and of hippocampal volume loss once corticosteroids were tapered and discontinued.^1,18 One study reported that >90% of patients recovered within 6 weeks, with patients with delirium recovering more quickly (mean: 5.4 days) than those with depression, mania, or psychosis (mean: 19.3 days).³ Another found that the vast majority (92%) of patients treated only with a steroid taper achieved clinical recovery, and 84% recovered with administration of antipsychotics without a steroid taper.³ In this study, all patients who received electroconvulsive therapy (ECT) recovered, as did those who received a steroid taper plus lithium or antipsychotics. Steroid tapering regimens are especially important for patients who have received long-term glucocorticoid treatment. Patients need to be closely monitored for signs of new or increased depression, delirium, or confusion during the taper. If these symptoms occur, the patient should be checked for adrenocortical insufficiency, which can be resolved by re-administering or increasing the dosage of the glucocorticoid.¹⁰

Mania. The treatment of mania/hypomania includes mood stabilizers (valproate, lithium, lamotrigine) and antipsychotics (quetiapine, olanzapine, haloperidol).^{2,4,5,10,14,18} Valproate has been reported to be an effective prophylactic of corticosteroid-induced mania,² perhaps because it dampens neuronal hyperexcitability by attenuating NMDA receptors, blocking voltage-dependent sodium channels, and inhibiting the synthesis of cortical GABAergic steroids. Starting valproate while continuing corticosteroids (if necessary) may help lessen mania.² Benzodiazepines also may be useful on a short-term basis. 

Depression. Steroid-induced depression may be treated with sertraline or other first-line antidepressants.^5,14 Consider ECT for patients with severe depression. Support for the use of antipsychotic medications stems from studies that reported steroids’ role in disrupting dopamine and 5HT2 activity. Lithium also has been used successfully to manage and prevent glucocorticoid-associated affective disorder.^10,18 It can be used alone or in combination with selective serotonin reuptake inhibitors to alleviate depressive symptoms.¹⁰ Tricyclic antidepressants are generally avoided because their anticholinergic effects can exacerbate or worsen delirium.¹⁸ In general, ECT is an effective treatment for persistent and/or unresponsive steroid-induced depression,^2,10 but may be difficult to use in patients with serious medical illnesses.

Agitation. Medications that have been proposed for treating steroid-induced agitation include benzodiazepines, haloperidol, and second-generation antipsychotics.^5,17

Other considerations. Clinicians, patients, and families should discuss in detail the risks of steroid-induced psychiatric symptoms so an early diagnosis and appropriate intervention can be implemented. Before starting steroids, it is important to review the patient’s current medication list to ensure that steroid treatment is indicated, and to check for potential drug–drug interactions. In addition, the medical condition that is being treated with steroids also needs to be carefully reviewed, because certain illnesses are associated with the development of psychiatric symptoms. ^5,10

Continue to: Young children...

Young children (age <6) and older adults appear to be at greater risk for cognitive and memory disturbances from steroid use.¹⁰ In addition, patients have individual levels of susceptibility to steroid-induced psychiatric symptoms that can vary over time. The risk for adverse effects may be elevated based on response to previous courses of glucocorticoid treatment.¹⁰ While gender, age, dosage, and duration of treatment influence risk, it is not possible to predict which patients will experience psychiatric effects during a given course of glucocorticoid therapy. Therefore, all patients should be considered to have the potential of developing such effects, and should be monitored during glucocorticoid treatment and withdrawal.

Goals for future research

To help reduce the severity of and cost associated with steroid-induced psychiatric symptoms,^5,14 future studies should focus on controlled trials of preventative strategies. In particular, recent advances in genetic mapping may help identify involvement of certain genes or polymorphisms.⁵ Because current guidelines for the prevention and treatment of steroid-induced psychiatric symptoms are not evidence-based, controlled clinical trials are needed to elucidate the optimal management of such symptoms. There is much interplay between many of the proposed mechanisms of steroid-induced psychiatric symptoms, and future studies can help uncover a deeper understanding of the intricacies of this phenomenon.

CASE CONTINUED

Mrs. N is admitted for altered mental status. Medical workup includes MRI of the brain, MRI of the neck, cardiac echocardiogram, and EEG. There is no evidence of acute structural pathology. She is started on olanzapine, 10 mg/d at bedtime for manic and psychotic symptoms, and is discharged after 5 days. After 1 month, the outpatient psychiatrist gradually decreases and discontinues olanzapine as Mrs. N steadily returns to baseline. One year after discharge, Mrs. N continues to report resolution of her manic and psychotic symptoms.

Bottom Line

Steroids can induce a wide range of psychiatric symptoms, including mania/ hypomania, anxiety, and depression. Initial treatment typically includes tapering or discontinuing the steroid when possible. Other proposed treatments include certain antipsychotics, antidepressants, and other psychotropics, but the supporting evidence is largely anecdotal or based on case studies. Additional research is needed to elucidate the mechanism and treatment recommendations.

Related Resources

• Janes M, Kuster S, Goldson TM, et al. Steroid-induced psychosis. Proc (Bayl Univ Med Cent). 2019;32(4):614-615.
• Mayo Clinic. Prednisone and other corticosteroids. https://www.mayoclinic.org/steroids/art-20045692

Drug Brand Names

Haloperidol • Haldol
Lamotrigine • Lamictal
Lithium • Eskalith, Lithobid
Methylprednisolone injection • Solu-Medrol
Olanzapine • Zyprexa
Paroxetine • Paxil
Phenytoin • Dilantin
Quetiapine • Seroquel
Risperidone • Risperdal
Sertraline • Zoloft
Valproate • Depakote

Ms. N, age 30, presents to the emergency department for altered mental status, insomnia, and behavioral changes, which she has experienced for 1 week. On evaluation, she grabs a clinician’s hand and details her business ideas and life story with no prompting. Ms. N’s mental status examination is significant for hyperverbal speech with increased rate and volume; tangential thought process; and bright, expanded affect.

One week earlier, Ms. N was hospitalized for sudden-onset chest pain, weakness, and dizziness. She received 45 minutes of cardiopulmonary resuscitation prior to presentation and was found to have a ST-segment elevation myocardial infarction that required emergent left anterior descending coronary artery and right coronary artery percutaneous coronary intervention to place drug-eluting stents. Her recovery was complicated by acute cardiogenic shock, pulmonary edema, and hypoxic respiratory failure. Subsequently, she was intubated, admitted to the ICU, and received high-dose corticosteroids, including IV methylprednisolone, 40 mg every 12 hours, which was tapered prior to discharge. Her husband reports that since Ms. N came home, she has been more talkative and irritable, ruminating about past events, unable to sleep (<1 hour/night), and crying frequently. She has also been endorsing visual and auditory hallucinations, with increased praying and listening to religious music.

The frequent clinical use of steroids necessitates an understanding of these medications’ various adverse effects. The manifestations of steroid-induced psychiatric symptoms are broad and can involve affective, behavioral, and cognitive domains. While the current mechanism is unknown, this phenomenon may be related to decreased levels of corticotropin, norepinephrine, and beta-endorphin immunoreactivity, as well as effects on brain regions such as the hippocampus and amygdala. The best interventions for steroid-induced psychiatric symptoms are awareness and early diagnosis. There are no FDA-approved treatments for steroid-induced psychiatric symptoms; initial measures should include tapering or discontinuing corticosteroids.

In this article, we review the literature on the incidence, characteristics, differential diagnoses, proposed mechanism, risk factors, and proposed treatments of steroid-induced psychiatric symptoms.

A wide range of presentations

Steroid use has increased over the past 2 decades, with 10% of medical and surgical inpatients and 1% to 3% of the general population taking long-term glucocorticoids.¹ Even with topical application, steroid therapy is often systemically absorbed, and thus may lead to steroid-induced psychiatric symptoms. The incidence of steroid-induced psychiatric symptoms is difficult to assess because there can be a wide range of reactions that are dose- and time-related. Three reviews of a total of 122 cases reports found that an estimated 5% of patients treated with steroids experience severe psychiatric reactions.^1-3

Steroid-induced psychopathology can include mood, behavioral, and/or cognitive impairments. Mania/hypomania is the most common overall psychiatric symptom; the most common mood manifestations are anxiety and depression.^4,5 Other possible steroid-induced symptoms include psychosis, dementia, panic disorder, delirium, suicidal thinking and behavior, aggressive behavior, insomnia, agitation, depersonalization, and euphoria.⁵ The most common cognitive impairment is verbal or declarative memory deficit; others include distractibility and deficits in attention and psychomotor speed.⁵ These psychiatric symptoms can have a rapid onset, possibly within hours of starting steroids.¹ However, studies have reported a median time to onset of 11.5 days; 39% of cases had onset during the first week and 62% within 2 weeks.^3,6 After reducing or stopping the steroid, it may take days to weeks before symptoms start to subside.²

What to consider in the differential Dx

Psychiatric symptoms that are induced by steroids can mimic metabolic, neurologic, or toxic disorders. Other factors to consider include drug withdrawal/intoxication, infections, and paraneoplastic syndromes.^4,5 Although there is no reported correlation between the location of neurologic lesions and the development of specific psychiatric symptoms, manic symptoms appear most commonly with lesions in the right frontal lobe. ⁴ Other factors to note include the presence of new-onset psychiatric illnesses such as bipolar, mood, or thought disorders,⁴ as well as psychosocial stressors that might be contributing to the patient’s presentation.⁵

Continue to: Proposed mechanisms

Although the exact mechanism by which steroids induce psychiatric symptoms is unknown, several mechanisms have been proposed. One hypothesis is that steroid-induced psychopathology is related to decreased levels of corticotropin, norepinephrine, and beta-endorphin immunoreactivity.^4,5,7 This may explain why many patients with major depressive disorder have elevated cortisol production and/or lack of suppression of cortisol secretion during a dexamethasone stimulation test, and why approximately one-half of patients with Cushing’s disease experience depressive symptoms.⁸ This is also likely why antipsychotics, which typically reduce cortisol, are efficacious treatments for some steroid-induced psychiatric symptoms.⁹ 

Cognitive impairments from steroid use may be related to these agents’ effects on certain brain regions. One such area is the hippocampus, an important mediator in the creation and maintenance of episodic and declarative memories.^5,8,9 Acute glucocorticoid use is associated with decreased activity in the left hippocampus, reduced hippocampal glucose metabolism, and reduced cerebral blood flow in the posterior medial temporal lobe.¹⁰ Long-term glucocorticoid exposure is associated with smaller hippocampal volume and lower levels of temporal lobe N-acetylaspartate, a marker of neuronal viability.¹⁰ Because working memory depends on the prefrontal cortex and declarative memory relies on the hippocampus, deficits in these functions can be attributed to the effect of prolonged glucocorticoid exposure on glucocorticoid or mineralocorticoid receptors in the hippocampus, reduction of hippocampal volume, or elevated glutamate accumulation in that area.¹¹ In addition, high cortisol levels inhibit brain-derived neurotrophic factor, which plays a crucial role in maintaining neural architecture in key brain regions such as the hippocampus and prefrontal cortex.¹¹ There is also a correlation between the duration of prednisone treatment and atrophy of the right amygdala, which is an important regulator of mood and anxiety.¹¹ Both the hippocampus and amygdala have dense collections of glucocorticoid receptors. This may explain why patients who receive high-dose corticosteroids can have reversible atrophy in the hypothalamus and amygdala, leading to deficits in emotional learning and the stress response.

Factors that increase risk

Several factors can increase the risk of steroid-induced psychopathology. The most significant is the dose; higher doses are more likely to produce psychiatric symptoms.^1,5 Concurrent use of drugs that increase circulating levels of corticosteroids, such as inhibitors of the cytochrome P450 (CYP) enzyme (eg, clarithromycin), also increases the likelihood of developing psychiatric symptoms.^1,5 Risk is also increased in patients with liver or renal dysfunction.¹ Cerebral spinal fluid/serum albumin ratio, a marker of blood-brain barrier damage, and low serum complement levels were also reported to be independent risk factors,¹² with the thought that increased permeability of the blood-brain barrier may allow hydrophobic steroid molecules to more easily penetrate the CNS, leading to increased neuropsychiatric effects. Hypoalbuminemia is another reported risk factor, perhaps because lower levels of serum albumin are related to higher levels of free and active glucocorticoids, which are normally inactive when bound to albumin.¹³ There also appears to be an increased prevalence of steroid-induced psychopathology in women, perhaps due to greater propensity in women to seek medical care or a higher prevalence of women with medical disorders that are treated with steroids.⁵ A previous history of psychiatric disorders may not increase risk.⁵

Several methods for reducing risk have been proposed, including using a divided-dosing regimens that may lower peak steroid plasma concentrations.^13,14 However, the best prevention of steroid-induced psychiatric symptoms are awareness, early diagnosis, and intervention. Studies have suggested that N-methyl-d-aspartate (NMDA) antagonists¹⁵ and other agents that decrease glutamate release (such as phenytoin and lamotrigine¹⁶) may help prevent corticosteroid-induced hippocampal volume loss. Lamotrigine has been shown to reduce the amount of atrophy in the amygdala in patients taking corticosteroids.¹⁷ Phenytoin has also been reported to reduce the incidence of hypomania associated with corticosteroids, perhaps due to its induction of CYP450 activity and acceleration of steroid clearance.¹⁶

Treatment options

There are no FDA-approved medications for managing steroid-induced psychiatric symptoms.^1,16 Treatment is based on evidence from case reports and a few small case series (Table^2-5,17,18).

Continue to: When possible, initial treatment...

When possible, initial treatment should include discontinuing or tapering corticosteroids to <40 mg/d of prednisone-equivalent.^1,4,10,18 Most studies have reported rapid reversal of deficits in declarative memory and of hippocampal volume loss once corticosteroids were tapered and discontinued.^1,18 One study reported that >90% of patients recovered within 6 weeks, with patients with delirium recovering more quickly (mean: 5.4 days) than those with depression, mania, or psychosis (mean: 19.3 days).³ Another found that the vast majority (92%) of patients treated only with a steroid taper achieved clinical recovery, and 84% recovered with administration of antipsychotics without a steroid taper.³ In this study, all patients who received electroconvulsive therapy (ECT) recovered, as did those who received a steroid taper plus lithium or antipsychotics. Steroid tapering regimens are especially important for patients who have received long-term glucocorticoid treatment. Patients need to be closely monitored for signs of new or increased depression, delirium, or confusion during the taper. If these symptoms occur, the patient should be checked for adrenocortical insufficiency, which can be resolved by re-administering or increasing the dosage of the glucocorticoid.¹⁰

Mania. The treatment of mania/hypomania includes mood stabilizers (valproate, lithium, lamotrigine) and antipsychotics (quetiapine, olanzapine, haloperidol).^{2,4,5,10,14,18} Valproate has been reported to be an effective prophylactic of corticosteroid-induced mania,² perhaps because it dampens neuronal hyperexcitability by attenuating NMDA receptors, blocking voltage-dependent sodium channels, and inhibiting the synthesis of cortical GABAergic steroids. Starting valproate while continuing corticosteroids (if necessary) may help lessen mania.² Benzodiazepines also may be useful on a short-term basis. 

Depression. Steroid-induced depression may be treated with sertraline or other first-line antidepressants.^5,14 Consider ECT for patients with severe depression. Support for the use of antipsychotic medications stems from studies that reported steroids’ role in disrupting dopamine and 5HT2 activity. Lithium also has been used successfully to manage and prevent glucocorticoid-associated affective disorder.^10,18 It can be used alone or in combination with selective serotonin reuptake inhibitors to alleviate depressive symptoms.¹⁰ Tricyclic antidepressants are generally avoided because their anticholinergic effects can exacerbate or worsen delirium.¹⁸ In general, ECT is an effective treatment for persistent and/or unresponsive steroid-induced depression,^2,10 but may be difficult to use in patients with serious medical illnesses.

Agitation. Medications that have been proposed for treating steroid-induced agitation include benzodiazepines, haloperidol, and second-generation antipsychotics.^5,17

Other considerations. Clinicians, patients, and families should discuss in detail the risks of steroid-induced psychiatric symptoms so an early diagnosis and appropriate intervention can be implemented. Before starting steroids, it is important to review the patient’s current medication list to ensure that steroid treatment is indicated, and to check for potential drug–drug interactions. In addition, the medical condition that is being treated with steroids also needs to be carefully reviewed, because certain illnesses are associated with the development of psychiatric symptoms. ^5,10

Continue to: Young children...

Young children (age <6) and older adults appear to be at greater risk for cognitive and memory disturbances from steroid use.¹⁰ In addition, patients have individual levels of susceptibility to steroid-induced psychiatric symptoms that can vary over time. The risk for adverse effects may be elevated based on response to previous courses of glucocorticoid treatment.¹⁰ While gender, age, dosage, and duration of treatment influence risk, it is not possible to predict which patients will experience psychiatric effects during a given course of glucocorticoid therapy. Therefore, all patients should be considered to have the potential of developing such effects, and should be monitored during glucocorticoid treatment and withdrawal.

Goals for future research

To help reduce the severity of and cost associated with steroid-induced psychiatric symptoms,^5,14 future studies should focus on controlled trials of preventative strategies. In particular, recent advances in genetic mapping may help identify involvement of certain genes or polymorphisms.⁵ Because current guidelines for the prevention and treatment of steroid-induced psychiatric symptoms are not evidence-based, controlled clinical trials are needed to elucidate the optimal management of such symptoms. There is much interplay between many of the proposed mechanisms of steroid-induced psychiatric symptoms, and future studies can help uncover a deeper understanding of the intricacies of this phenomenon.

CASE CONTINUED

Mrs. N is admitted for altered mental status. Medical workup includes MRI of the brain, MRI of the neck, cardiac echocardiogram, and EEG. There is no evidence of acute structural pathology. She is started on olanzapine, 10 mg/d at bedtime for manic and psychotic symptoms, and is discharged after 5 days. After 1 month, the outpatient psychiatrist gradually decreases and discontinues olanzapine as Mrs. N steadily returns to baseline. One year after discharge, Mrs. N continues to report resolution of her manic and psychotic symptoms.

Bottom Line

Steroids can induce a wide range of psychiatric symptoms, including mania/ hypomania, anxiety, and depression. Initial treatment typically includes tapering or discontinuing the steroid when possible. Other proposed treatments include certain antipsychotics, antidepressants, and other psychotropics, but the supporting evidence is largely anecdotal or based on case studies. Additional research is needed to elucidate the mechanism and treatment recommendations.

Related Resources

• Janes M, Kuster S, Goldson TM, et al. Steroid-induced psychosis. Proc (Bayl Univ Med Cent). 2019;32(4):614-615.
• Mayo Clinic. Prednisone and other corticosteroids. https://www.mayoclinic.org/steroids/art-20045692

Drug Brand Names

Haloperidol • Haldol
Lamotrigine • Lamictal
Lithium • Eskalith, Lithobid
Methylprednisolone injection • Solu-Medrol
Olanzapine • Zyprexa
Paroxetine • Paxil
Phenytoin • Dilantin
Quetiapine • Seroquel
Risperidone • Risperdal
Sertraline • Zoloft
Valproate • Depakote

Ms. N, age 30, presents to the emergency department for altered mental status, insomnia, and behavioral changes, which she has experienced for 1 week. On evaluation, she grabs a clinician’s hand and details her business ideas and life story with no prompting. Ms. N’s mental status examination is significant for hyperverbal speech with increased rate and volume; tangential thought process; and bright, expanded affect.

One week earlier, Ms. N was hospitalized for sudden-onset chest pain, weakness, and dizziness. She received 45 minutes of cardiopulmonary resuscitation prior to presentation and was found to have a ST-segment elevation myocardial infarction that required emergent left anterior descending coronary artery and right coronary artery percutaneous coronary intervention to place drug-eluting stents. Her recovery was complicated by acute cardiogenic shock, pulmonary edema, and hypoxic respiratory failure. Subsequently, she was intubated, admitted to the ICU, and received high-dose corticosteroids, including IV methylprednisolone, 40 mg every 12 hours, which was tapered prior to discharge. Her husband reports that since Ms. N came home, she has been more talkative and irritable, ruminating about past events, unable to sleep (<1 hour/night), and crying frequently. She has also been endorsing visual and auditory hallucinations, with increased praying and listening to religious music.

The frequent clinical use of steroids necessitates an understanding of these medications’ various adverse effects. The manifestations of steroid-induced psychiatric symptoms are broad and can involve affective, behavioral, and cognitive domains. While the current mechanism is unknown, this phenomenon may be related to decreased levels of corticotropin, norepinephrine, and beta-endorphin immunoreactivity, as well as effects on brain regions such as the hippocampus and amygdala. The best interventions for steroid-induced psychiatric symptoms are awareness and early diagnosis. There are no FDA-approved treatments for steroid-induced psychiatric symptoms; initial measures should include tapering or discontinuing corticosteroids.

In this article, we review the literature on the incidence, characteristics, differential diagnoses, proposed mechanism, risk factors, and proposed treatments of steroid-induced psychiatric symptoms.

A wide range of presentations

Steroid use has increased over the past 2 decades, with 10% of medical and surgical inpatients and 1% to 3% of the general population taking long-term glucocorticoids.¹ Even with topical application, steroid therapy is often systemically absorbed, and thus may lead to steroid-induced psychiatric symptoms. The incidence of steroid-induced psychiatric symptoms is difficult to assess because there can be a wide range of reactions that are dose- and time-related. Three reviews of a total of 122 cases reports found that an estimated 5% of patients treated with steroids experience severe psychiatric reactions.^1-3

Steroid-induced psychopathology can include mood, behavioral, and/or cognitive impairments. Mania/hypomania is the most common overall psychiatric symptom; the most common mood manifestations are anxiety and depression.^4,5 Other possible steroid-induced symptoms include psychosis, dementia, panic disorder, delirium, suicidal thinking and behavior, aggressive behavior, insomnia, agitation, depersonalization, and euphoria.⁵ The most common cognitive impairment is verbal or declarative memory deficit; others include distractibility and deficits in attention and psychomotor speed.⁵ These psychiatric symptoms can have a rapid onset, possibly within hours of starting steroids.¹ However, studies have reported a median time to onset of 11.5 days; 39% of cases had onset during the first week and 62% within 2 weeks.^3,6 After reducing or stopping the steroid, it may take days to weeks before symptoms start to subside.²

What to consider in the differential Dx

Psychiatric symptoms that are induced by steroids can mimic metabolic, neurologic, or toxic disorders. Other factors to consider include drug withdrawal/intoxication, infections, and paraneoplastic syndromes.^4,5 Although there is no reported correlation between the location of neurologic lesions and the development of specific psychiatric symptoms, manic symptoms appear most commonly with lesions in the right frontal lobe. ⁴ Other factors to note include the presence of new-onset psychiatric illnesses such as bipolar, mood, or thought disorders,⁴ as well as psychosocial stressors that might be contributing to the patient’s presentation.⁵

Continue to: Proposed mechanisms

Although the exact mechanism by which steroids induce psychiatric symptoms is unknown, several mechanisms have been proposed. One hypothesis is that steroid-induced psychopathology is related to decreased levels of corticotropin, norepinephrine, and beta-endorphin immunoreactivity.^4,5,7 This may explain why many patients with major depressive disorder have elevated cortisol production and/or lack of suppression of cortisol secretion during a dexamethasone stimulation test, and why approximately one-half of patients with Cushing’s disease experience depressive symptoms.⁸ This is also likely why antipsychotics, which typically reduce cortisol, are efficacious treatments for some steroid-induced psychiatric symptoms.⁹ 

Cognitive impairments from steroid use may be related to these agents’ effects on certain brain regions. One such area is the hippocampus, an important mediator in the creation and maintenance of episodic and declarative memories.^5,8,9 Acute glucocorticoid use is associated with decreased activity in the left hippocampus, reduced hippocampal glucose metabolism, and reduced cerebral blood flow in the posterior medial temporal lobe.¹⁰ Long-term glucocorticoid exposure is associated with smaller hippocampal volume and lower levels of temporal lobe N-acetylaspartate, a marker of neuronal viability.¹⁰ Because working memory depends on the prefrontal cortex and declarative memory relies on the hippocampus, deficits in these functions can be attributed to the effect of prolonged glucocorticoid exposure on glucocorticoid or mineralocorticoid receptors in the hippocampus, reduction of hippocampal volume, or elevated glutamate accumulation in that area.¹¹ In addition, high cortisol levels inhibit brain-derived neurotrophic factor, which plays a crucial role in maintaining neural architecture in key brain regions such as the hippocampus and prefrontal cortex.¹¹ There is also a correlation between the duration of prednisone treatment and atrophy of the right amygdala, which is an important regulator of mood and anxiety.¹¹ Both the hippocampus and amygdala have dense collections of glucocorticoid receptors. This may explain why patients who receive high-dose corticosteroids can have reversible atrophy in the hypothalamus and amygdala, leading to deficits in emotional learning and the stress response.

Factors that increase risk

Several factors can increase the risk of steroid-induced psychopathology. The most significant is the dose; higher doses are more likely to produce psychiatric symptoms.^1,5 Concurrent use of drugs that increase circulating levels of corticosteroids, such as inhibitors of the cytochrome P450 (CYP) enzyme (eg, clarithromycin), also increases the likelihood of developing psychiatric symptoms.^1,5 Risk is also increased in patients with liver or renal dysfunction.¹ Cerebral spinal fluid/serum albumin ratio, a marker of blood-brain barrier damage, and low serum complement levels were also reported to be independent risk factors,¹² with the thought that increased permeability of the blood-brain barrier may allow hydrophobic steroid molecules to more easily penetrate the CNS, leading to increased neuropsychiatric effects. Hypoalbuminemia is another reported risk factor, perhaps because lower levels of serum albumin are related to higher levels of free and active glucocorticoids, which are normally inactive when bound to albumin.¹³ There also appears to be an increased prevalence of steroid-induced psychopathology in women, perhaps due to greater propensity in women to seek medical care or a higher prevalence of women with medical disorders that are treated with steroids.⁵ A previous history of psychiatric disorders may not increase risk.⁵

Several methods for reducing risk have been proposed, including using a divided-dosing regimens that may lower peak steroid plasma concentrations.^13,14 However, the best prevention of steroid-induced psychiatric symptoms are awareness, early diagnosis, and intervention. Studies have suggested that N-methyl-d-aspartate (NMDA) antagonists¹⁵ and other agents that decrease glutamate release (such as phenytoin and lamotrigine¹⁶) may help prevent corticosteroid-induced hippocampal volume loss. Lamotrigine has been shown to reduce the amount of atrophy in the amygdala in patients taking corticosteroids.¹⁷ Phenytoin has also been reported to reduce the incidence of hypomania associated with corticosteroids, perhaps due to its induction of CYP450 activity and acceleration of steroid clearance.¹⁶

Treatment options

There are no FDA-approved medications for managing steroid-induced psychiatric symptoms.^1,16 Treatment is based on evidence from case reports and a few small case series (Table^2-5,17,18).

Continue to: When possible, initial treatment...

When possible, initial treatment should include discontinuing or tapering corticosteroids to <40 mg/d of prednisone-equivalent.^1,4,10,18 Most studies have reported rapid reversal of deficits in declarative memory and of hippocampal volume loss once corticosteroids were tapered and discontinued.^1,18 One study reported that >90% of patients recovered within 6 weeks, with patients with delirium recovering more quickly (mean: 5.4 days) than those with depression, mania, or psychosis (mean: 19.3 days).³ Another found that the vast majority (92%) of patients treated only with a steroid taper achieved clinical recovery, and 84% recovered with administration of antipsychotics without a steroid taper.³ In this study, all patients who received electroconvulsive therapy (ECT) recovered, as did those who received a steroid taper plus lithium or antipsychotics. Steroid tapering regimens are especially important for patients who have received long-term glucocorticoid treatment. Patients need to be closely monitored for signs of new or increased depression, delirium, or confusion during the taper. If these symptoms occur, the patient should be checked for adrenocortical insufficiency, which can be resolved by re-administering or increasing the dosage of the glucocorticoid.¹⁰

Mania. The treatment of mania/hypomania includes mood stabilizers (valproate, lithium, lamotrigine) and antipsychotics (quetiapine, olanzapine, haloperidol).^{2,4,5,10,14,18} Valproate has been reported to be an effective prophylactic of corticosteroid-induced mania,² perhaps because it dampens neuronal hyperexcitability by attenuating NMDA receptors, blocking voltage-dependent sodium channels, and inhibiting the synthesis of cortical GABAergic steroids. Starting valproate while continuing corticosteroids (if necessary) may help lessen mania.² Benzodiazepines also may be useful on a short-term basis. 

Depression. Steroid-induced depression may be treated with sertraline or other first-line antidepressants.^5,14 Consider ECT for patients with severe depression. Support for the use of antipsychotic medications stems from studies that reported steroids’ role in disrupting dopamine and 5HT2 activity. Lithium also has been used successfully to manage and prevent glucocorticoid-associated affective disorder.^10,18 It can be used alone or in combination with selective serotonin reuptake inhibitors to alleviate depressive symptoms.¹⁰ Tricyclic antidepressants are generally avoided because their anticholinergic effects can exacerbate or worsen delirium.¹⁸ In general, ECT is an effective treatment for persistent and/or unresponsive steroid-induced depression,^2,10 but may be difficult to use in patients with serious medical illnesses.

Agitation. Medications that have been proposed for treating steroid-induced agitation include benzodiazepines, haloperidol, and second-generation antipsychotics.^5,17

Other considerations. Clinicians, patients, and families should discuss in detail the risks of steroid-induced psychiatric symptoms so an early diagnosis and appropriate intervention can be implemented. Before starting steroids, it is important to review the patient’s current medication list to ensure that steroid treatment is indicated, and to check for potential drug–drug interactions. In addition, the medical condition that is being treated with steroids also needs to be carefully reviewed, because certain illnesses are associated with the development of psychiatric symptoms. ^5,10

Continue to: Young children...

Young children (age <6) and older adults appear to be at greater risk for cognitive and memory disturbances from steroid use.¹⁰ In addition, patients have individual levels of susceptibility to steroid-induced psychiatric symptoms that can vary over time. The risk for adverse effects may be elevated based on response to previous courses of glucocorticoid treatment.¹⁰ While gender, age, dosage, and duration of treatment influence risk, it is not possible to predict which patients will experience psychiatric effects during a given course of glucocorticoid therapy. Therefore, all patients should be considered to have the potential of developing such effects, and should be monitored during glucocorticoid treatment and withdrawal.

Goals for future research

To help reduce the severity of and cost associated with steroid-induced psychiatric symptoms,^5,14 future studies should focus on controlled trials of preventative strategies. In particular, recent advances in genetic mapping may help identify involvement of certain genes or polymorphisms.⁵ Because current guidelines for the prevention and treatment of steroid-induced psychiatric symptoms are not evidence-based, controlled clinical trials are needed to elucidate the optimal management of such symptoms. There is much interplay between many of the proposed mechanisms of steroid-induced psychiatric symptoms, and future studies can help uncover a deeper understanding of the intricacies of this phenomenon.

CASE CONTINUED

Mrs. N is admitted for altered mental status. Medical workup includes MRI of the brain, MRI of the neck, cardiac echocardiogram, and EEG. There is no evidence of acute structural pathology. She is started on olanzapine, 10 mg/d at bedtime for manic and psychotic symptoms, and is discharged after 5 days. After 1 month, the outpatient psychiatrist gradually decreases and discontinues olanzapine as Mrs. N steadily returns to baseline. One year after discharge, Mrs. N continues to report resolution of her manic and psychotic symptoms.

Bottom Line

Steroids can induce a wide range of psychiatric symptoms, including mania/ hypomania, anxiety, and depression. Initial treatment typically includes tapering or discontinuing the steroid when possible. Other proposed treatments include certain antipsychotics, antidepressants, and other psychotropics, but the supporting evidence is largely anecdotal or based on case studies. Additional research is needed to elucidate the mechanism and treatment recommendations.

Related Resources

• Janes M, Kuster S, Goldson TM, et al. Steroid-induced psychosis. Proc (Bayl Univ Med Cent). 2019;32(4):614-615.
• Mayo Clinic. Prednisone and other corticosteroids. https://www.mayoclinic.org/steroids/art-20045692

Drug Brand Names

Haloperidol • Haldol
Lamotrigine • Lamictal
Lithium • Eskalith, Lithobid
Methylprednisolone injection • Solu-Medrol
Olanzapine • Zyprexa
Paroxetine • Paxil
Phenytoin • Dilantin
Quetiapine • Seroquel
Risperidone • Risperdal
Sertraline • Zoloft
Valproate • Depakote