Psychosocial and Psychopharmacological Interventions for Comorbid Bipolar Disorder and Substance Use Disorders: A Review

Abstract

The co-occurrence of bipolar disorder and substance use disorders presents significant clinical challenges, complicating diagnosis, treatment, and prognosis. This review explores both psychosocial and psychopharmacological interventions evaluated in randomized, controlled trials to address this comorbidity. Findings indicate that while several interventions show promise, no single treatment is consistently effective across both mood and substance use domains as compared to placebo. These findings highlight the necessity of personalized, multimodal treatment plans tailored to the specific needs of individuals with comorbid bipolar disorder and substance use disorders.

Bipolar disorder (BD) is a condition characterized by shifting periods of elevated mood (also known as mania or hypomania) and depressed mood. The burden of these episodes is often worsened by the presence of co-occurring psychiatric conditions, such as substance use disorders (SUDs), that can occur along with BD. The comorbidity of BD and SUDs is significant, ranging from ~17% to 40% across studies.¹ Comorbid BD and SUDs can lead to decreased recovery rates from mood episodes in BD, reduced quality of life, increased hospitalization rates, and higher rates of attempted suicide.² The significant illness burden associated with comorbid BD and SUDs suggests the value of specialized treatments that can target the unique clinical picture of comorbid BD and SUDs. This article explores psychosocial and psychopharmacological treatments that have been evaluated for the treatment of comorbid BD and SUDs in the context of randomized, controlled trials. We conclude with a synthesis of these evaluated treatments and discuss future directions for the treatment and research of comorbid BD and SUDs.

PSYCHOSOCIAL INTERVENTIONS FOR COMORBID BD AND SUDS

Integrated Group Therapy

Integrated group therapy (IGT) is the most studied psychosocial approach for the treatment of comorbid BD and SUDs.^3–5 IGT stems from the idea that the skills required for recovery and preventing relapse are relevant for both disorders and their treatment, thus emphasizing the interaction of BD and SUD.³ Group topics include denial, ambivalence, and acceptance, managing mood symptoms without turning to substance use, and identifying and fighting triggers for mood episodes and substance misuse.³

Weiss and colleagues (2007) compared 20 weeks of IGT to 20 weeks of group drug counseling (GDC), with a 3-month follow-up period.⁵ GDC includes 20 weekly hour-long sessions and it targets only substance abuse. Each GDC session begins with a check-in about substance use in the past week and continues with substance-use-related modules that do not target mood issues (eg, facilitating abstinence or coping with substance-related problems). Eligible participants for this study were 18 years old or older, had a current diagnosis of BD and SUD, endorsed substance use within the past 60 days, and reported being on a mood-stabilizing regimen for more than 2 weeks. Participants (N = 62) were randomly assigned to IGT (n = 31) or GDC (n = 31). The primary mood outcome was the number of weeks in a mood episode and the primary substance abuse outcome was the days of substance use.⁵ Posttreatment, the days of use for any substance decreased significantly more in IGT compared to GDC (P < 0.001). Additionally, during follow-up, substance use remained lower in IGT compared to GDC (P < 0.001). No significant between-group differences existed in the number of weeks in a mood episode.

Weiss and colleagues (2009) replicated the study using a shorter version of IGT (12 treatment sessions rather than 20 treatment sessions, with a 3-month follow-up period).⁴ This version of IGT included more educational information about cognitive-behavioral therapy (CBT) and BD to support clinicians who did not have a background in these areas. Participants (N = 61) were randomly assigned to IGT (n = 31) or GDC (n = 30). The primary outcomes were identical to the 2007 trial. IGT was more effective in improvements in substance use symptoms and in abstinence compared to GDC, but this difference was not significant during treatment or follow-up. There were also no significant between-group differences in mood outcomes.⁴

Integrated Cognitive-Behavioral Treatments

In a randomized controlled trial, Schmitz and colleagues compared an integrated 12-week medication management (MM) and individual CBT treatment to MM only for patients aged 18–55 with comorbid BD and SUDs.⁶ CBT included sixteen 60-minute sessions and focused on targeting comorbid BD and SUD through education and skills training (eg, daily self-monitoring of mood, drug craving, and medication adherence). Medication management consists of four 20-minute sessions in which the participant answered questions about medication adherence, medication side effects, mood, and substance use. Participants (N = 46) were randomly assigned to MM (n = 21) or MM+CBT (n = 25).⁶ The primary outcomes included drug use (via self-report or urine drug test), self-reported alcohol use, and mood symptoms (days of depressive and manic symptoms). There were statistically significant improvements in depression (P < 0.05) and mania (P < 0.01) in the MM+CBT group compared to the MM group. There were no between-group differences in alcohol use outcomes.⁶

Integrated Treatment Adherence Program

In a randomized controlled trial, Wenze and colleagues compared a 6-month Integrated Treatment Adherence Program (ITAP) to treatment as usual for participants aged 18 years and older.⁷ ITAP is based on identifying participants’ personal values, such that participants set goals and choose behaviors that align with those values. ITAP includes three 60-minute individual in-person sessions, a 60-minute in-person family session with the study “significant other” (ie, friend, family member) and 11 brief (15–30 minutes) phone check-ins with the participant and their significant other, separately. The in-person sessions focus on psychoeducation about BD and SUD, discussion of personal values, and clarification of life goals. Treatment as usual typically involved medication management. Participants (N = 39) were randomly assigned to receive ITAP (n = 20) or treatment as usual (n = 19).⁷ The primary outcomes were self-reported depressive symptoms, manic symptoms, and daily drug and alcohol use. Participants randomized to ITAP saw faster and greater improvement in their depression and mania scores compared to participants in treatment as usual (P < 0.05). There were no significant between-group differences in substance use at posttreatment follow-up.⁷

PSYCHOPHARMACOLOGICAL INTERVENTIONS FOR COMORBID BD AND SUDS

Valproic Acid

In a 24-week double-blind placebo-controlled randomized parallel-group trial of add-on valproic acid or placebo to ongoing lithium for the treatment of comorbid bipolar I disorder (BDI) with active alcohol use disorder, participants (ages 18–65) were randomly assigned to receive valproate (n = 27) or placebo (n = 25). Valproate dosage was titrated to a target trough serum concentration of 50 to 100 μg/mL.⁸

The primary alcohol use outcomes included proportion of heavy drinking days (defined as ≥ 4 standard drinks per day for women and ≥ 5 standard drinks per day for men), time to relapse to sustained heavy drinking (defined as 3 consecutive heavy drinking days), remission of mania, and remission of depression. Valproate significantly reduced the proportion of heavy drinking days at week 24 relative to placebo (44% vs 68%, P = 0.02). Additionally, valproate treatment extended the time to relapse to sustained heavy drinking to a median of 75 days, compared to 44 days in the placebo group (P = 0.048). There was no significant between-group difference in manic or depressive symptoms. Both saw a substantial decrease in manic symptoms, with average mania scores dropping by about 60%. There were no serious drug-related adverse events, although nausea and vomiting were more common in the valproate group.⁸

Quetiapine

Three studies have investigated quetiapine in individuals with comorbid BD and alcohol use disorder.^9–11 A 12-week randomized double-blind placebo-controlled trial evaluated add-on quetiapine in patients with comorbid BD and alcohol use disorders.¹¹ Participants (ages 18–55, N = 102) were randomized to receive quetiapine (titrated up to 600 mg/day, n = 52) or placebo (n = 50). There was no significant difference between the two groups in alcohol use, alcohol craving, or mood outcomes, though the quetiapine group did have a statistically significant initial decline in depression severity in the first half of the study (P = 0.04).

These investigators then conducted a similar study among participants with BD (mixed or depressed phase) and comorbid alcohol use disorder, but with higher baseline alcohol use.¹⁰ This trial was a 12-week randomized double-blind placebo-controlled trial of sustained release quetiapine (up to 600 mg/day, n = 44) vs placebo (n = 44). Participants were 18–65 years of age, had consumed at least 15 drinks in the 7 days prior to baseline, and were on a stable dose of mood-stabilizing medication for ≥ 14 days. At posttreatment follow-up, there were again no significant differences in mood or substance use outcomes. Quetiapine was relatively well tolerated.

A larger (N = 362) 12-week double-blind placebo-controlled study of quetiapine (flexible dosing up to 800 mg, n = 176) vs placebo (n = 186) added to lithium or divalproex in participants with BDI and alcohol use disorder included participants who reported heavy drinking for at least 10 of the 28 days prior to their screening session.⁹ The primary outcome was the change in the proportion of heavy drinking days from baseline to week 12 with mood symptoms assessed as secondary outcomes. There were no statistically significant differences between groups in substance or in mood-related outcomes. The reported adverse events were consistent with the established side-effect profile of quetiapine.

Naltrexone

This study investigated the efficacy of naltrexone as an add-on to therapy for alcohol use disorder in a 12-week randomized double-blind parallel-group placebo-controlled trial among individuals with BD and alcohol dependence.¹² Participants (ages 18–70, N = 43) were randomly assigned to naltrexone (fixed at 50 mg/day; n = 20) or placebo (n = 23), consumed at least 5 drinks in the 7 days prior to the intake, and were currently experiencing depressed or mixed symptoms. All participants also received manual-driven CBT designed for comorbid BD and alcohol use disorders. There were no significant between-group differences in mood (depression, mania) or alcohol use outcomes.¹²

Acamprosate

An 8-week randomized double-blind placebo-controlled clinical trial evaluated the effects of acamprosate (1,998 mg/day; n = 16) vs placebo (n = 17) among individuals with comorbid BD and alcohol use disorder.¹³ Overall, no statistically significant differences were seen in alcohol use outcomes or mood outcomes. However, post hoc analysis indicated that participants in the acamprosate group had significantly lower clinical global impression (CGI) scores of substance use severity at weeks 7 and 8 compared to baseline (P < 0.05). Acamprosate was well tolerated.¹³

Citicoline

Brown and colleagues (2007) conducted a randomized controlled parallel-group proof-of-concept trial to evaluate the efficacy of citicoline (target dose 2,000 mg/day, n = 44) vs placebo (n = 23) as an add-on therapy to improve cognition in a 12-week study of patients with BD and comorbid cocaine use disorder.¹⁴ Secondary outcomes were the likelihood of a cocaine-positive urine screen at posttreatment follow-up and mood symptoms (depression, mania). The placebo group had more than 6 times the odds of testing positive for cocaine at study exit compared to the citicoline group (P = 0.026). There were no significant between-group differences in mood symptoms. Citicoline was well tolerated.¹⁴

Brown and colleagues (2015) conducted an additional randomized double-blind parallel-group placebo-controlled study of citicoline for comorbid BDI and cocaine use disorder over a 12-week period.¹⁵ Eligible participants were in a depressed or mixed-mood state, treated with mood-stabilizing medication, and had cocaine use within 7 days of baseline. Participants (N = 122) were randomly assigned to receive either citicoline (n = 61) or placebo (n = 61). The primary outcome was the presence or absence of a cocaine-positive urine screen at posttreatment follow-up and secondary outcomes evaluated symptoms of depression or mania. There were significant differences in this primary outcome, with the citicoline group less likely to have a cocaine-positive urine screen at posttreatment follow-up (P = 0.022). There were no significant differences in mood outcomes. Citicoline was well tolerated.¹⁵

Brown and colleagues (2019) also conducted a randomized double-blind parallel-group placebo-controlled study of citicoline for individuals with BD and comorbid alcohol use disorder over a 12-week period.¹⁶ Participants (N = 55) were randomly assigned to receive either citicoline (n = 29) or placebo (n = 26). The primary outcomes were changes in alcohol consumption (ie, number of heavy drinking days) and cravings, with depressive symptoms assessed as a secondary outcome. There were no significant differences in the mood or substance use outcomes between the citicoline and placebo groups. Again, citicoline was well tolerated.¹⁶

Lamotrigine

This study was a randomized double-blind placebo-controlled trial comparing lamotrigine therapy to placebo among patients with BD and comorbid cocaine use disorder over a 10-week period.¹⁷ Participants (N = 112) were randomly assigned to lamotrigine (maximum dose of 400 mg/day, n = 55) or placebo (n = 57) added to existing psychiatric medications. There was no significant between-group difference in substance use outcomes. However, the dollar amount spent on cocaine was statistically lower for the lamotrigine group (P = 0.05). There were no significant differences in depressive or manic symptoms between groups. Lamotrigine was well tolerated.¹⁷

DISCUSSION

In this review, we evaluated psychosocial and psychopharmacological interventions that have been explored to date for the treatment of comorbid BD and SUDs in the context of randomized, controlled trials. Across the psychosocial trials, the treatment arms were generally more effective relative to control arms, but typically not for both mood and substance use outcomes. This observation also aligns with findings from a prior review evaluating psychosocial interventions for comorbid bipolar disorder and substance use disorders.¹⁸ For example, among the psychosocial interventions, IGT was associated with significant improvements in substance use symptoms, but not in mood symptoms,^4,5 MM+CBT was associated with significant improvements in mood symptoms, but not in substance use symptoms,⁶ and ITAP was associated with significant improvements in mood symptoms, but not in substance use symptoms.⁷ Across the psychopharmacological trials, unlike in the psychosocial trials, treatment arms were not consistently more effective than the control arms. Similar to the psychosocial trials, when treatment arms were more effective, they were generally not effective for both mood and substance use outcomes. For example, both acamprosate and citicoline were associated with significant improvements in the substance use domain, but not in the mood symptom domain relative to placebo^13–15 (though, in one study, citicoline was not associated with significant improvements in mood or substance use outcomes relative to placebo¹⁶). Quetiapine, naltrexone, and lamotrigine were not associated with significant improvements in mood or substance use symptoms relative to placebo. In summary, we observe that, across psychosocial and psychopharmacological studies, when treatments were effective, they generally proved to be better for mood or substance use in this comorbid population, but not both.

This review is limited by the small number of randomized, controlled studies in the psychosocial domain. It is also possible that findings could be impacted by the type of SUD. For example, the citicoline studies that had a positive mood effect were in cocaine use disorder, whereas the study with null findings was in alcohol use disorder. Overall, these findings reflect the challenges associated with identifying a singular treatment that can yield improvements across the symptom domains (ie, mood, substance use) that are relevant to patients with comorbid BD and SUDs. These findings may suggest the importance of personalized medicine approaches for the treatment of comorbid BD and SUDs moving forward. Specifically, rather than identifying a single treatment that can cover the range of mood and substance use symptoms, it may be most helpful to select a variety of treatments that are balanced among a given patient’s symptom presentation, treatment preference, treatment history, and associated demographic and clinical characteristics.