Bipolar disorder and substance abuse: Overcome the challenges of ‘dual diagnosis’ patients
After testing positive for cocaine on a recent court-mandated urine drug screen, Mr. M, age 49, is referred by his parole officer for psychiatric and substance abuse treatment. Mr. M has bipolar I disorder and alcohol, cocaine, and opioid dependence. He says he has been hospitalized or incarcerated at least once each year for the past 22 years. Mr. M has seen numerous psychiatrists as an outpatient, but rarely for more than 2 to 3 months and has not taken any psychotropics for more than 5 months.
Approximately 1 week before his recent urine drug screen, Mr. M became euphoric, stopped sleeping for several days, and spent $2,000 on cocaine and “escorts.” He reports that each day he smokes 30 cigarettes and drinks 1 pint of liquor and prefers to use cocaine and opiates by IV injection. Several years ago Mr. M was diagnosed with hepatitis C virus (HCV) but received no further workup or treatment. Although he denies manic or psychotic symptoms, Mr. M is observed speaking to unseen others in the waiting room and has difficulty remaining still during his interview. His chief concern is insomnia, stating, “Doc, I just need something to help me sleep.”
The high prevalence of substance use disorders (SUDs) in persons with bipolar disorder (BD) is well documented.1,2 Up to 60% of bipolar patients develop an SUD at some point in their lives.3-5 Alcohol use disorders are particularly common among BD patients, with a lifetime prevalence of roughly 50%.2-5 Recent epidemiologic data indicate that 38% of persons with bipolar I disorder and 19% of those with bipolar II disorder meet criteria for alcohol dependence.5 Comorbid SUDs in patients with BD are associated with:
- poor treatment compliance
- longer and more frequent mood episodes
- more mixed episodes
- more hospitalizations
- more frequent suicide attempts.1,2
The impact of co-occurring SUDs on suicidality is particularly high among those with bipolar I disorder.6
Frequently referred to as “dual diagnosis” conditions, co-occurring BD and SUDs may be more accurately envisioned as multi-morbid, rather than comorbid, illnesses. Data from the Stanley Foundation Bipolar Network suggest that 42% of BD patients have a lifetime history of ≥2 comorbid axis I disorders.7 Rates of generalized anxiety disorder, panic disorder, and posttraumatic stress disorder are particularly high in BD patients with co-occurring SUDs.8,9 In addition, the presence of 1 SUD may mark the presence of other SUDs; for example, alcohol dependence is strongly associated with polysubstance abuse, especially in females with BD.8 Furthermore, medical comorbidities that impact treatment decisions also are highly prevalent in BD patients with comorbid SUDs.10 In particular, HCV rates are higher in persons with BD compared with the general population,11 and are >5 times as likely in bipolar patients with co-occurring SUDs.12
Unfortunately, limited treatment research guides clinical management of comorbid BD and SUDs.2,13,14 Clinical trials of medications for BD traditionally have excluded patients with SUDs, and persons with serious mental illness usually are ineligible for SUD treatment studies.13 Furthermore, the few randomized controlled trials (RCTs) conducted in persons with both illnesses have been constrained by relatively small sample sizes and low retention rates. In the absence of a definitive consensus for optimal treatment, this article outlines general clinical considerations and an integrated approach to assessing and managing this complex patient population.
Birds of a feather
Multiple hypotheses try to account for the high rate of SUDs in patients with BD (Table 1), but none fully explain the complex interaction observed clinically.14,15 Substance dependence and BD are chronic remitting/relapsing disorders with heterogeneous presentations and highly variable natural histories. As with SUDs, BD may go undiagnosed and untreated for years, coming to clinical attention only after substantial disease progression.16
The fluctuating illness course in BD and SUDs makes diagnosis and treatment difficult. Symptomatic periods often are interrupted by spontaneous remissions and longer—although usually temporary—periods of perceived control. Both BD and SUDs may be associated with profound mood instability, increased impulsivity, altered responsiveness to reward, and impaired executive function.17 Finally, the high degree of heritability in BD18 and many SUDs19 may make treatment engagement more difficult if either disorder is present in multiple family members because of:
- potential for greater clinical severity
- reduced psychosocial resources
- altered familial behavioral norms that may impede the patient’s recognition of illness.
Denial of illness is a critical symptom that may fluctuate with disease course in both disorders. Persons with BD or SUDs may be least likely to recognize that they are ill during periods of highest symptom severity. Accordingly, treatment adherence in patients with either disorder may be limited at baseline and decline further when the 2 illnesses co-occur.20
Involvement in the criminal justice system and medical comorbidities, particularly HCV, also complicate diagnosis and treatment of BD patients with SUDs. For more information about these topics, see Box1 and Box2.
Why is substance abuse so prevalent among bipolar patients?
Selected limitations of this hypothesis
Self-medication: substance abuse occurs as an attempt to regulate mood
High rates of substance use during euthymia; high prevalence of alcohol/depressant use during depressive phase, stimulant use during manic phase
Common neurobiologic or genetic risk factors
Specific evidence from linkage/association studies currently is lacking
Substance use occurs as a symptom of bipolar disorder
High percentage of patients with bipolar disorder do not have SUDs; there is a poor correlation of onset, course of bipolar, and SUD symptoms
Substance use unmasks bipolar disorder or a bipolar diathesis
Emergence of mania before SUD is common and predictive of more severe course of bipolar disorder
High comorbidity rates are an artifact of misdiagnosis based on overlapping symptoms and poor diagnostic boundaries
Very high prevalence of SUDs also is observed in longitudinal studies of patients initially hospitalized for mania
SUDs: substance use disorders
Integrated clinical management
Assessment. Although not intended to be comprehensive, suggestions for routine assessment of patients with suspected SUDs and/or BD are listed in Table 2. Because clinicians may encounter dual diagnosis patients in general psychiatric clinics or specialty (addiction or mood disorder) clinics, it is useful to obtain a thorough substance use history in all patients with known or suspected BD as well as a thorough history of hypomania/mania and depression in all patients with addictive disorders. BD diagnoses by self-report or chart history in patients with SUDs should be considered cautiously because BD often is overdiagnosed in persons engaged in active substance abuse or experiencing withdrawal.21 If past or present mood symptoms and substance use have co-occurred, further focused assessment of mood symptoms before alcohol and drug use or during extended periods of abstinence are necessary to make the diagnosis of bipolar disorder with confidence. Family history of SUDs and/or BD are neither necessary nor sufficient for either diagnosis; however, collateral information from family or significant others could help make the diagnosis and may identify aids and obstacles to treatment planning and engagement.
When a patient’s clinical history strongly supports the diagnoses of BD and co-occurring SUDs, more detailed inquiry is warranted. Determining the patient’s age at onset of each disorder may have prognostic value because onset of mania before SUDs developed, especially in adolescence, may predict a more severe course of both illnesses.22 A complete alcohol use history should include routine questioning about past withdrawal. Previous withdrawal seizures in an actively drinking BD patient may tip the balance toward adding an anticonvulsant for mood stabilization. A thorough SUD history should elicit information about polysubstance abuse or dependence and include screening for injection drug use and other risk factors for HCV and human immunodeficiency virus (HIV), such as hypersexuality during manic or hypomanic episodes. Document the date of the last screening for HCV/ HIV in BD patients at high risk of infection. The U.S. Centers for Disease Control and Prevention recommends that all patients at high risk for HIV consider voluntary screening at least annually.23
Assess your patient’s historical and ongoing alcohol and other drug abuse at the initial visit, and continue to monitor substance use at all subsequent visits, especially in patients with HCV. When feasible, order urine drug screening and laboratory testing for alcohol use biomarkers such as carbohydrate-deficient transferrin and gamma-glutamyltransferase to supplement self-report data, especially in patients with poor insight or low motivation. Assess suicidal ideation and any changes in suicide risk factors at every visit.
Treatment. No biologic therapies have been FDA-approved for treating patients with co-occurring BD and SUDs. Comorbid SUDs in BD patients—as well as rapid cycling and mixed mood episodes, both of which are more common in patients with comorbid SUDs—predict poor response to lithium.17 However, the evidence base for optimal pharmacotherapy remains extremely limited. Published double-blind, placebo-controlled RCTs in persons with BD and co-morbid SUDs are limited to only 1 trial each of lithium, carbamazepine, quetiapine, and naltrexone, and 2 comparisons of lithium plus divalproex vs lithium alone (Table 3).24-29
Salloum et al24 reported that bipolar I disorder patients with alcohol dependence who received divalproex plus lithium as maintenance treatment had fewer heavy drinking days and fewer drinks per heavy drinking day than those receiving lithium plus placebo. However, the addition of divalproex did not improve manic or depressive symptoms, and depression remission rates remained relatively low in both groups. A recent 6-month study comparing lithium vs lithium plus divalproex in patients with SUDs and rapid-cycling BD found no additional benefit of divalproex over lithium monotherapy in retention, mood, or substance use outcomes.28 However, modest evidence that anticonvulsants such as valproic acid and carbamazepine may help treat acute alcohol withdrawal2 could support their preferential use as mood stabilizers over lithium in actively drinking BD patients.
Research underscores the difficulty in keeping dual diagnosis patients in treatment. Salloum et al24 reported that only one-third of randomized subjects completed the 24-week study. In the Kemp study,28 79% of 149 recruited subjects failed to complete the lead-in stabilization phase. Of the 31 remaining subjects who were randomized to lithium or lithium/divalproex combination, only 8 (26% of those randomized, 5% of those recruited) completed the 6-month trial.
Substance abuse is associated with significantly decreased treatment adherence in persons with BD20 and may affect medication choice. For example, caution may be warranted in the use of lamotrigine in substance-abusing patients with poor adherence because re-titration from the starting dose is recommended if the medicine has been missed for a consecutive period exceeding 5 half-lives of the drug.30
Notable progress has been made in developing psychosocial treatments for comorbid SUDs and BD. Integrated group therapy has been designed to address the 2 disorders simultaneously by emphasizing the relationship between the disorders and highlighting similarities in cognitive and behavioral change that promote recovery in both.31 In a recent well-controlled RCT, this approach reduced alcohol and other drug use to approximately one-half the levels observed in those who received only group drug counseling.31
Research suggests that an integrated approach that encompasses psychiatric, medical, psychosocial, and legal dimensions simultaneously may be most effective. For patients such as Mr. M, this would include aggressively treating mood symptoms while employing motivational interviewing techniques to improve engagement in substance dependence treatment. If possible, involving family members, parole officials, housing agencies, and other public assistance workers in the treatment plan may increase treatment adherence and reduce loss of contact during illness exacerbations. Stabilization of substance use and psychiatric morbidity should be accompanied by timely evaluation of HCV and other medical comorbidities in order to improve long-term prognosis.
Assessing patients you suspect have comorbid BD and SUDs
Thorough substance use history in all patients with known or suspected bipolar disorder:
Thorough evaluation of any history of hypomania/mania and depression in all patients with known or suspected SUDs:
Assess risk factors, screening status for hepatitis C, HIV
Obtain collateral information from family and significant others if feasible and appropriate
Detailed assessment of suicide risk
Substance use since last visit by self-report
Consider UDS, CDT, GGT
Detailed assessment of suicide risk
BD: bipolar disorder; CDT: carbohydrate-deficient transferrin; GGT: gamma-glutamyltransferase; HIV: human immunodeficiency virus; SUDs: substance use disorders; UDS: urine drug screen
Pharmacotherapy for bipolar disorder and co-occurring SUDs
Salloum et al, 200524
BD I, alcohol dependence. N=59 
Divalproex + lithium vs lithium, 24 weeks
Decreased number of heavy drinking days, fewer drinks per heavy drinking day
Geller et al, 199825
BD I, BP II, MDD (adolescents); alcohol, cannabis abuse. N=25 
Lithium, 6 weeks
Decreased cannabis-positive urine drug screen (lithium > placebo)
Brady et al, 200226
BD I, BD II, cyclothymia; cocaine dependence. N=57
Carbamazepine, 12 weeks
Trend toward longer time to cocaine use
Brown et al, 200827
BD I, BD II; N=102
Quetiapine, 12 weeks
Decreased HAM-D scores (quetiapine > placebo)
Kemp et al, 200928
BD I, BD II, rapid cycling; alcohol, cannabis, cocaine abuse or dependence. N=31 
Divalproex + lithium vs lithium, 6 months
No group differences
Brown et al, 200929
BD I, BD II; alcohol dependence. N=50 
Naltrexone, 12 weeks
Trend toward increased probability of no drinking days
* N=number of subjects randomized to double-blind treatment. Numbers in brackets indicate the number of subjects who completed all study visits (when reported)
BD: bipolar disorder; HAM-D: 17-item Hamilton Rating Scale for Depression; MDD: major depressive disorder; SUDs: substance use disorders