An estimated 1–2 % of older adults in the general population have alcohol use disorders (AUD) [54]. Specific estimates of geriatric bipolar disorder patients with this SUD range from 13 to 25 %, making it the most common substance use disorder [12]. Research on the relationship between alcohol use disorder severity and suicide risk has produced mixed results [46]. Carra [46] reported that both alcohol dependence and alcohol abuse were associated with increased risk of suicide attempts.

Guidelines for the workup of alcohol use disorders among geriatric bipolar patients are described in Substance Abuse among Older Adults. Treatment Improvement Protocol (TIP) Series 26 published by SAMHS [55]. Of special note is the Michigan Alcoholism Screening Test-Geriatric Version, which has been adapted for use in this population for other substances and validated for different ethnic groups [56].

Since patients with comorbid psychiatric disorders are routinely excluded from clinical trials of new treatments for SUD, the safety and efficacy of existing treatments for alcohol use and other SUDs are emerging areas of investigation for patients with bipolar disorder [57]. Given that adults over age 60 are also often excluded from these studies, there is even less information about effective, safe treatment options for older adults with the bipolar disorder–SUD comorbidity. Furthermore, our understanding of psychiatric and medical comorbidity in bipolar disorder derives largely from studies of younger and mixed-age samples [12]. However, it is a clinical priority to identify more effective evidence-based pharmacological [58] and behavioral treatments for comorbid AUD and bipolar disorder across age groups given the role of alcohol dependence in contributing to occupational dysfunction in bipolar disorder through a variety of mechanisms [59] Addressing substance abuse may beneficially impact the course of bipolar disorder. Improved stabilization of bipolar disorder is also likely to aid substance disorder treatment compliance [8].

Patients with AUD and a co-occurring psychiatric disorder have more complex symptoms than patients with either disorder alone, are least likely to engage in psychosocial interventions, and benefit the most from a full arsenal of treatment options including psychotherapy as well as pharmacotherapy [60, 61]. Although the cumulative scope and impact of bipolar disorder with co-occurring alcohol use disorder across age groups have been well documented, there are little data on selection and implementation of optimal treatment strategies [62, 63]. None of the current FDA approved treatments for alcohol use (i.e., disulfiram, naltrexone, and acamprosate) have been demonstrated to be clearly effective among patients with bipolar disorder [57]. Valproate [64] and the combination of lithium and valproate have been shown to improve mood symptoms as well as reduce days of heavy drinking [65]. Several studies suggest that quetiapine lacks efficacy in the treatment of comorbid bipolar disorder and AUD [66, 67]. Other studies of treatment with acamprosate and naltrexone have also failed to show a statistically significant benefit in patients with bipolar disorder and AUD [68, 69].

Several investigators suggest that topiramate is effective in reducing heavy drinking among individuals with alcohol dependence [22, 70]. However, a randomized, placebo-controlled trial of topiramate for mixed-age adults with bipolar disorder and alcohol dependence (compared with a general population of alcohol dependent patients) (N = 12) did not support adjunctive use of topiramate [71].

Behavioral therapies for alcohol dependence—including 12-step programs, cognitive behavioral therapy for relapse prevention, and motivational interviewing—merit further exploration, including through randomized controlled studies, for older adults with bipolar disorder. In one mixed methods study of dually diagnosed patients, including OABD patients, the peer support group and peer education emphases of 12-step programs such as Alcoholics Anonymous were identified as helpful even by patients with significant psychiatric symptom burden (i.e., while still experiencing some level of psychosis or mania) [72]. Pilot studies of cognitive behavioral therapy for relapse prevention (including “Third Wave” cognitive behavioral approaches such as mindfulness-based relapse prevention) have shown significant benefit for older adults, including veterans, though have not been specifically tested in older adults with bipolar disorder and comorbid alcohol dependence [73]. Similarly, while case series reports on using integrated motivational interviewing and cognitive behavioral therapies have demonstrated promise among bipolar disorder patients (across age groups) with comorbid AUDs, there is a lack of randomized controlled trials of these psychosocial treatments for OABD patients. Potential future randomized clinical trials could apply existing knowledge to the OABD population, such as from protocols developed by Schmitz and Grillion [74] and Weiss et al. [75], including in integrated community-based group therapy.

Substance use disorders involving benzodiazepines as well as other sedative-hypnotics (such as zolpidem, eszopiclone, and zaleplon) are common despite the well-known adverse effects of these medications among elderly patients (including increased fall risk and cognitive impairment). The estimated prevalence of benzodiazepine use in the general geriatric patient population is 11 % based on limited studies [76], making this SUD in some settings (e.g., outpatient psychiatry) nearly as common as alcohol use disorders. The risk of long-term continuous sedative-hypnotic use (not recommended by most prescribing guidelines) is greater among elderly patients, reflecting greater rates of anxiety and insomnia within this group [77]. It is not yet clear what role gender and ethnicity may play in prevalence. However, Petrovic et al. [78] and colleagues found that female gender, widowed status, borderline personality disorder, and anxiety disorders were risk factors for benzodiazepine dependence, and immigrant patients from developing countries may be at particular risk given high population-wide use (up to 26.7 % among women over age 60 and 14 % among males) of long-acting benzodiazepines for extended duration in Latin America, for example [79]. While studies to identify risk factors within older adults with bipolar disorder are still lacking, bipolar disorder is a known risk factor for benzodiazepine dependence in general due to the high rate of anxiety disorder comorbidity, with nearly 50 % of adults with BD meeting anxiety disorders criteria per lifetime. Further, some raise the possibility that anxiety symptoms, which are often treated and self-treated with sedative-hypnotics, may be primary features of bipolar illness rather than a comorbidity [80].

In addition to the SAMHSA guidelines referenced above, tools such as the A-FRAMES (Assessment, Feedback, Responsibility, Advice, Menu, Empathy, and Self-efficacy) Questionnaire modify existing motivational interviewing tools to address geriatric issues [81]. Other tools are being developed for home evaluations, for more sensitive detection, and for use among prescribers (e.g., STOPP—Screening Tool of Older Person’s Potentially Inappropriate Prescriptions) in inpatient and outpatient settings [82].

Pharmacotherapy rests on use of benzodiazepines in detoxification taper protocol, with potential adjunctive medications proposed such as acetylcholinesterase inhibitors [83], phenobarbital [84], and oxcarbazepine [85]. These medications have yet to be tested among geriatric bipolar disorder patients in treatment for chronic benzodiazepine or other sedative-hypnotic use but hold relevant potential given wide baseline use for mood and cognition within the population. As for other older adults or other groups of patients with advanced liver disease, including cirrhosis, benzodiazepines that do not get oxidized by the liver (oxazepam and lorazepam) are preferable to chlordiazepoxide to minimize the risk of oversedation, accumulation, or toxicity, also given that oxazepam and lorazepam are shorter acting [86]. It is also worth noting that, contrary to the practice (in the 1980s) of “only using alprazolam to detoxify patients from alprazolam,” this motto is not supported by clinical evidence, and the recommendation is to use either oxazepam or lorazepam among elderly patients, with phenobarbital substitution also a possibility. Falls, myocardial infarctions, and delirium without signs of autonomic hyperactivity are risks of benzodiazepine withdrawal in the elderly general population, to which dually diagnosed patients may be particularly vulnerable given polypharmacy; thus, medically supervised benzodiazepine withdrawal among OABD patients is critical [87].

Stimulant-induced psychosis or mania can mimic affective disorder-induced episodes in those without bipolar disorder. Among dual diagnosis patients, stimulants may also exacerbate actual mood episodes. Thus, it is critical to consider these substance use disorders (SUDs) on the differential diagnosis given high medical sequelae potentially more severe in elderly populations: primarily cardiovascular and neurologic, including tachycardia and other arrhythmias, cerebrovascular accidents (CVA), seizures (for which elderly patients may be at additional risk already due to polypharmacy), cocaine- and amphetamine-induced myocardial ischemia and infarction, and infectious diseases [88].

Notably, over 80 % of cocaine users, even those without comorbid bipolar disorder as a risk factor, experience paranoia. Further, upwards of 55 % in one cohort reported cocaine-related violent behaviors [89]. Both homicide and suicide risk are elevated in relation to the length and severity of cocaine use. Psychiatric symptoms are more pronounced among crack cocaine users than those dependent on other forms of cocaine (i.e., intranasal powdered cocaine) [89]. Male gender appears to be predictive of cocaine use in the elderly population, though studies specific to geriatric bipolar patients are still lacking. One emergency department study, conducted over a 6-month period of ED visits made by persons above the age of 60, demonstrated that 2 % of these individuals had cocaine-positive urines, with an average age of 66 and 90 % male [90].

Diagnosis by routine toxicology may prove difficult given the short half-life of cocaine (remaining in urine 4–8 h after the dose, with benzoylecgonine (the psychoactive metabolite of cocaine) potentially remaining in urine for up to 60 h after use) and trends toward episodic or binge rather than daily use [91]. Specialized screening tools are under investigation, including adaptation of the Decision Balance Questionnaire for assessing vulnerability to cocaine relapse [92], although none have been validated for geriatric bipolar patients to date.

Pharmacological studies designed for bipolar disorder patients with coexisting cocaine use disorder have shown initial promising results, including for citicoline, lamotrigine, and quetiapine in recent trials. Other drugs that have been considered in case reports, but not yet tested in randomized clinical studies, include gabapentin, mifepristone, valproate, and pramipexole. A randomized placebo-controlled trial of the cognitive enhancer citicoline, a moderator of phospholipid metabolism that increases acetylcholine levels, demonstrated improved declarative memory and an increased abstinence rate from cocaine, although citicoline did not improve mood [93]. In a 10-week randomized controlled trial of lamotrigine among bipolar disorder patients (depressed or in a mixed episode) with cocaine dependence, self-reported cocaine use (measured by dollars spent) decreased significantly. A follow-up uncontrolled replication study showed improvements in measures of cocaine craving, mood, and drug use [94], while an earlier open-label study also showed a significant decrease in craving [95]. A more recent study showed a similar reduction in cocaine use with benefits occurring earlier in treatment and attenuating over time [96]. Notably, drug dreams, predictors of use and correlated with research study treatment survival among patients with SUDs, also appeared to decrease in other larger studies of lamotrigine [97]. There is less evidence of quetiapine’s efficacy for the bipolar cocaine dependence comorbidity, with additional concerns of the possible abuse liability of this drug, though an open-label study (N = 17) showed good tolerability and improvement of psychiatric symptoms [98].

The rates of amphetamine use disorders (including prescribed psychostimulants, as well as illicit drugs such as methamphetamine) are unknown for geriatric bipolar patients specifically. However, given overlap between core affective disorder symptoms (such as impulsivity, risk taking, and euphoria) and the intoxication syndrome for amphetamines, as well as elevated rates of attention-deficit–hyperactivity disorders (ADHD) within the population, bipolar disorder patients in general are at particular risk [99]. Of the estimated 5 % of adults in the general population, over 20 % may have comorbid bipolar disorder [100].

While use of stimulants within the population carries a risk of mania, as well as the medical complications discussed above for cocaine (including cardiovascular and neurologic), studies are also underway of stimulant augmentation of mood stabilizers to address bipolar depression, with preliminary promising results [101, 102].

Prevalence of stimulant dependence estimated from a voluntary bipolar disorder patient registry (N = 100) was 2.8 % among those with bipolar I patients, with an age range of 18–65 and a specific prevalence unavailable for patients over age 60 [103]. While specific trends by ethnicity or gender are not known for geriatric bipolar patients, it is notable that bipolar disorder (along with anxiety disorders) was highly correlated with stimulant dependence in a group of gay and transgendered (GBT) men seeking outpatient psychiatric treatment, making elderly LGBT populations a potential priority for outreach and assessment [104].

In addition to a thorough medical workup, including electrocardiogram and urine drug screen (which remain positive for amphetamines for 1–2 days after use), screening through Physical Symptom Screening Triggers checklists (as described in the SAMHSA Tip 26) that assess sleep changes, anxiety, agitation, and availability of stimulants in the household (i.e., prescribed to a family member) may be of particular relevance [1].

A limited evidence base exists for adjunctive treatments of stimulant dependence among older adults with bipolar disorder. Nevertheless, effective mood stabilization remains the central strategy, and lithium’s neuroprotective benefits may ameliorate stimulant-induced reduction in brain choline levels [105]. Among behavioral treatments, in addition to those based on motivational interviewing, 12-step programs, and dual diagnosis models, recent research on contingency management (where participants earn vouchers or other compensation for continued program participation and/or continued verified negative urine screens) has been shown to decrease stimulant use among both younger and older bipolar disorder patients in community health settings [106]. Pharmacotherapy showing initial promise includes citicoline [107] (a randomized controlled trial demonstrating improved mood symptoms among mixed-age patients) and atypical antipsychotic medications. In a trial of quetiapine and risperidone among bipolar disorder patients of age 20–50, patients reported decreased drug craving related to decreased drug use [108]. Research among older adults with BD and psychostimulant use is warranted.

The most commonly abused prescription opioids by general population adults (across age group) are hydrocodone and oxycodone, although both have less abuse potential than the more potent mu agonists, morphine or fentanyl [109]. In large epidemiologic studies such as the National Epidemiologic Study on Alcohol and Related Conditions, there is evidence for bidirectional relationships between bipolar disorder and non-medical use of prescription opioids. Prescription opioid use appears to be a precipitating factor in bipolar mood episodes, the presentation of bipolar disorder, as well as a frequent complication of previously diagnosed bipolar disorder [110]. Male gender and young age are risk factors for prescription opioid use in the population [111], with a 28 % prevalence of bipolar disorder among male patients in one inpatient cohort, and mania a key risk of opioid detoxification [112, 113]. While men appear to be more likely to misuse prescription opioids, women are more likely to receive combination regimens of prescription opioids and sedative-hypnotics, placing them at possibly greater risk of overdose [114]. Other risk factors for prescription opioid use disorders include a positive family history of substance use, as well as personal exposure to violence and sexual assault, with some studies confirming that the diagnosis of bipolar illness constitutes a risk factor for subsequent trauma exposure for both men and women [115]. While comorbidity studies in older adults with bipolar disorder have been limited, rates of prescription opioid overdose deaths increased between 2000 and 2014 for patients over age 55 [116], with polypharmacy, medical comorbidities, and treatment refractory pain among the risk factors for mortality [117]. Up to 10 % of geriatric medical outpatients screen positive for bipolar disorder, type I or II, in cohorts receiving some form of an opioid pain regimen. Therefore, the increased risk of suicide and impulsive or accidental overdose that may be associated with bipolar illness merits thorough assessment for prescription opioid misuse among older adults with bipolar disorder [118].