Comorbidities in Youth with Bipolar Disorder: Clinical Features and Pharmacological Management

Abstract

Background: Bipolar Disorder (BD) is a highly comorbid condition, and rates of cooccurring disorders are even higher in youth. Comorbid disorders strongly affect clinical presentation, natural course, prognosis, and treatment.

Methods: This review focuses on the clinical and treatment implications of the comorbidity between BD and Attention-Deficit/Hyperactivity Disorder, disruptive behavior disorders (Oppositional Defiant Disorder and/or Conduct Disorder), alcohol and substance use disorders, Autism Spectrum Disorder, anxiety disorders, Obsessive-Compulsive Disorder, and eating disorders.

Results: These associations define specific conditions which are not simply a sum of different clinical pictures, but occur as distinct and complex combinations with specific developmental pathways over time and selective therapeutic requirements. Pharmacological treatments can improve these clinical pictures by addressing the comorbid conditions, though the same treatments may also worsen BD by inducing manic or depressive switches.

Conclusion: The timely identification of BD comorbidities may have relevant clinical implications in terms of symptomatology, course, treatment and outcome. Specific studies addressing the pharmacological management of BD and comorbidities are still scarce, and information is particularly lacking in children and adolescents; for this reason, the present review also included studies conducted on adult samples. Developmentally-sensitive controlled clinical trials are thus warranted to improve the prognosis of these highly complex patients, requiring timely and finely personalized therapies.

1. INTRODUCTION

Although diagnostic criteria for Bipolar Disorder (BD) provided by the Diagnostic and Statistical Manual of Mental Disorders (DSM) can be extended to youth, as indicated in the multisite, collaborative study on BD, the Course and Outcome of Bipolar Illness in Youth (COBY) [1], clinical phenotypes and boundaries in youth are still debated, given the possible developmentally different presentation of early-onset forms, compared to adult standards, including the rates and patterns of comorbidity [2]. Comorbidity, indeed, is the rule rather than the exception in early-onset BD, particularly with Attention-Deficit/Hyperactivity Disorder (ADHD), impulse control disorders - including disruptive behavior disorders (DBD) and substance use disorders (SUD) - anxiety disorders, Obsessive-Compulsive Disorder (OCD), Autism Spectrum Disorder (ASD) and other neurodevelopmental conditions [3-7]. A developmentally based approach to comorbidities may allow to explore the co-occurrence of disorders according to developmental trajectories to bipolarity, using both retrospective and prospective study designs [8].

Furthermore, early-onset impaired regulation of emotional states, including low threshold, rapid increase, excessive intensity, inappropriate expression, and slow normalization of emotionality, resulting in mood lability and instability, severe irritability, low tolerance to frustration, temper outburst, and hyper-arousal, is currently defined as Emotional Dysregulation [9-11], and is related to the cyclothymic temperament [12]. This condition can be interpreted as a neurodevelopmental disorder of emotion regulation, and a transdiagnostic prodrome of different mental disorders, including not only mood disorders, but also ADHD [13, 14], Oppositional Defiant Disorder (ODD) and Conduct Disorder (CD) [15, 16], anxiety disorders [17, 18], OCD [5, 19], ASD [20, 21] and eating disorders (ED) [22], resulting in individuals with more complex and unstable clinical expression, and increased sensitivity and vulnerability to environmental and interpersonal stimuli [23].

Among the different comorbidities, we will address here seven partially overlapping clinical conditions, with related treatment implications: a) BD and ADHD, b) BD and DBD including ODD/CD; c) BD and SUD; d) BD and ASD; e) BD and anxiety disorders; f) BD and OCD; g) BD and ED. The present review will mainly focus on the pharmacological treatments for BD comorbidities. Several other therapeutic options, including psychosocial interventions, are also available, though they will not be extensively discussed here; for a review on this topic, the authors may refer to [24-27].

2. MATERIALS AND METHODS

A narrative review of the literature was conducted. Due to the broad extent of the topic and the wide range of comorbid conditions considered in the article, a systematic review of the literature was not feasible. Moreover, given the paucity of controlled studies available for youth samples, it was also necessary to take into account relevant adult studies that could help inform clinicians dealing with children and adolescents. Studies were identified through multiple searches on PubMed from date of inception to December 2021. First, search terms included ‘bipolar disorder’ cross-referenced with ‘children’ or ‘adolescents’, along with the specific comorbid condition (e.g., ‘ADHD’). Secondly, ‘treatment’ was added to the search line. Finally, adult studies were retrieved including only ‘bipolar disorder’, ‘treatment’, and the specific comorbid condition in the search line. Previous systematic and narrative reviews were also considered eligible for revision. Reference lists from all identified publications were carefully searched for additional relevant studies.

3. ATTENTION-DEFICIT/HYPERACTIVITY DISORDER

ADHD and BD are highly prevalent conditions that share a chronic course and lead to significant impairment in educational, occupational and social functioning [28-30]. They often co-occur [30, 31], particularly in the combined presentation of ADHD [32]. Indeed, several studies reported high rates of ADHD comorbidity in pediatric BD, ranging from 11% to 98%, with an average rate of 62% [33]. At later ages, lower but still rather high rates of ADHD have been also reported in up to 23 - 25% of clinically referred adult patients with BD, though most studies consistently account for a comorbidity rate between 15% and 20% [34-36]. Across the age groups, ADHD is three times more prevalent among individuals with mood disorders compared to individuals without mood disorders, while among people with mood disorders, ADHD is 1.7 times more common in people with BD compared to people with Major Depressive Disorder [37].

Interestingly, ADHD has been found to increase the risk of transition from unipolar depression to BD [37, 38] and, according to a recent meta-analysis of prospective studies conducted by our group, more than 10% of ADHD patients are expected to be later diagnosed with BD [7]. Even more impressively, a nine-fold increase in the risk to develop BD can be observed in ADHD patients compared with healthy controls [7]. These findings are consistent with the five-to-ten-fold increase of risk found in patients with ADHD compared to healthy controls in nation-wide register-based studies [39,40]. Interestingly, among ADHD patients, emotional dysregulation has been found to predict greater rates of subsequent diagnoses of BD, possibly representing a specific manifestation pertaining to the bipolar spectrum in neurodevelopmental disorders [9].

Importantly, comorbid ADHD-BD patients exhibit higher severity of core ADHD symptoms, greater rates of externalizing disorders and addictive behaviors, including substance abuse, poorer educational functioning, more psychiatric hospitalizations, and higher rates of additional psychopathology [32, 35, 41-44]. They also show a more severe and chronic course of mood disorder with shorter euthymic intervals and more frequent mood episodes, higher rates of atypical and mixed depressions, greater affective instability and behavioral impulsivity, more non-suicidal self-injurious behaviors and suicide attempts [45, 46], and an overall reduced quality of life [44, 47-51]. Finally, a younger age of BD onset in ADHD patients compared to non-ADHD ones, with a mean difference of approximately five years, is one of the most replicated findings, and, among ADHD-BD patients, more than a half showed a pre-adult onset of BD (< 18 years) and more than a fourth a pre-pubertal onset (< 12 years) [50, 52].

Recent genome-wide association studies have implicated shared genetic factors across ADHD and BD suggesting a common genetic vulnerability, though shared environmental factors may also contribute to the high rates, including maltreatment, poverty and social adversities [53]. However, a recent population-based study found that there was no evidence for cross-transmission of BD and ADHD in first degree relatives, which may suggest that the comorbidity between ADHD and BD could instead represent a distinct BD subtype with childhood-onset symptoms - an “early-onset type” of BD [54].

Despite the high prevalence of ADHD-BD comorbidity and the particularly great severity of this condition [44], there are very few clinical trials addressing pharmacological interventions for ADHD in the BD population, the few existing trials being limited by small sample size and short-term follow-up. Since comorbid ADHD in BD is associated with negative outcomes both in children and adults [36, 55], a prompt and effective treatment is essential. Nonetheless, ADHD frequently goes untreated in people with mood disorders as it is often misrecognized or underestimated [29]. Indeed, individuals with comorbid ADHD and mood disorders may not receive proper treatment for ADHD due to clinicians’ focus on mood symptoms. This is supported by a previous investigation which found that only 9% of patients with BD and comorbid ADHD had been properly diagnosed with and received treatment for ADHD [56]. Furthermore, the comorbidity is associated with reduced ADHD treatment responsiveness and lower chance to recovery [42], while higher rates of antidepressant-induced switch, poorer response to mood stabilizers and less treatment adherence are among the major treatment challenges of the clinical management of these patients [44, 57, 58].

Preliminary results suggest that psychostimulants and other ADHD medications may be particularly effective in treating ADHD in people with mood disorders, while future research should be aimed to explore different treatment options, as well as acceptability, safety and efficacy of treating ADHD along with mood symptoms [59]. As for the effectiveness in the adult population, the Canadian Network for Mood and Anxiety Treatments (CANMAT) group drew a ranking based on levels of evidence and Methylphenidate (MPH) and mixed amphetamine salts (MAS) received, respectively, levels 1 and 2 of evidence for the treatment of ADHD in adults with comorbid BD. Level 3 was attributed to Bupropion and Atomoxetine (ATX), while level 4 to Modafinil, Venlafaxine, Desipramine, Nortriptyline and Lisdexamphetamine [59].

In children and adolescents, MPH may be considered as an effective treatment for comorbid ADHD-BD. Findings from the Multimodal Treatment Study of Children suggest that children with ADHD and BD respond robustly to MPH during the first month of treatment without significantly adverse effects [60]. Indeed, ADHD children with manic symptoms - identified using the Diagnostic Interview Schedule for Children or the Child Behavior Checklist as proxies of pediatric BD - showed an equivalent response to a one-month MPH titration trial as compared to ADHD peers without BD [60].

Similarly, pediatric patients with BD and concurrent ADHD can also be effectively treated with MAS after manic symptoms are stabilized. An 8-week open-label trial of Divalproex Sodium followed b-y a 4-week randomized, double-blind, placebo-controlled crossover trial with MAS has been conducted on 30 youths aged 6 to 17 years [61]. The authors found that, while Divalproex Sodium alone was not an effective treatment for ADHD in the context of BD, MAS was instead significantly more effective than placebo for ADHD core symptoms, with no significant side effects or worsening of manic symptoms [61]. More recently, other authors, based on case reports, proposed that ADHD-BD may be effectively treated with stimulants - e.g., MAS - even without mood-stabilizing agents [62].

Though, limited supporting evidence exists for non-stimulant medications. A consecutive retrospective case series on seven patients with comorbid ADHD-BD, who were treated with ATX and - all but one - mood stabilizers, demonstrated significant improvements in core ADHD symptoms; main adverse effects included sedation, nausea, and decreased appetite [63]. The same authors later conducted an open-label prospective study on twelve 6 to 14 years old patients with comorbid ADHD-BD with ongoing mood-stabilizers, and found that ATX was effective in improving ADHD core symptoms and behaviors but not BD symptoms [64]. ATX may thus be prescribed in individuals with ADHD and comorbid BD, though limited supporting evidence is available so far [65, 66]. Furthermore, the use of a combination of Amantadine and either Clonidine or Propranolol has been suggested to control symptoms of comorbid ADHD-BD by a recent case series conducted on three pediatric patients [67].

One of the major issues to consider when treating comorbid children and adolescents is the risk of mood switch. Individuals with comorbid ADHD-BD, particularly those with BD type 1, are at high risk for mood destabilization with most ADHD treatments and should thus be prescribed mood-stabilizing medications prior to initiate specific ADHD therapies [68]. Indeed, a previous retrospective study found an association between a past history of treatment with stimulants and an earlier onset of BD in adolescents with ADHD, suggesting that stimulants could accelerate the onset of BD [69]. Similarly, a retrospective study found that the odds of ADHD youths developing BD are significantly and positively associated with longer treatment with MPH and MAS prior to first diagnosis of mania [42].

Nevertheless, more recent encouraging evidence is available for MPH. In the Swedish National Registry, more than two thousand BD patients, who initiated a treatment with either MPH alone or with concomitant mood-stabilizers, were followed-up for 12 months [70]. The authors found that the risk of manic switch was initially increased in those patients assuming MPH monotherapy though at a non-significant level, and was unchanged in the successive period. By the contrary, in patients assuming both MPH and mood stabilizers, the risk was reduced in the first 3 months and unchanged in the successive period. Besides, stimulants do not seem to increase the risk of suicidal and non-suicidal self-injurious behavior in patients with ADHD, but, rather, may reduce such risk [71-73]. Similar, though limited findings are available also for patients with comorbid ADHD-BD. For instance, within six months after stimulants initiation, the rate of suicide attempts and non-suicidal self-injury events decreased significantly, both in terms of numbers of suicidal patients and numbers of events experienced; notably, these effects were preserved two years later [74]. Stimulant treatment may thus reduce the risk of suicidal events in patients with ADHD-BD. Thus, based on available studies, clinicians should not withhold stimulant treatment from ADHD patients for fear of deterioration of the concomitant BD, though mood stabilizer treatment and close monitoring remain warranted to minimize the risk of manic switches.

On the other hand, mixed evidence is available for the association between ATX and risk of both manic switch and suicidal ideation. Manic switch in ADHD patients without a prior history of psychotic illness or mania can be caused by ATX, though it is more common in patients with comorbid BD [75]. In a first retrospective chart review of seven pediatric patients with ADHD-BD already treated with mood stabilizers, no hypomania or mania was reported with the use of ATX for up to 18 months [63], while a worsening of mood symptoms was found in a minority of patients in an open-label prospective study [64]. In an Italian study on 781 young people, ATX induced a hypomanic episode only in one exposed patient [76], whereas in a Taiwanese registry study, short-term treatment with ATX in 145 children with ADHD did not correlate with a higher risk of BD onset [77]. An association between ATX and increased risk of suicidality has not been reported in ADHD patients, being such risk mostly related to comorbid conditions such as BD [64]; nevertheless, all patients treated with ATX should be carefully monitored for the emergence or worsening of suicidality [75].

Finally, since emotional dysregulation in ADHD patients has been found to possibly represent a specific predictor of subsequent diagnoses of BD [9], another major issue in the treatment of ADHD-BD comorbidity would be to directly address emotional dysregulation as a core manifestation pertaining to the bipolar spectrum in ADHD. Recently, a meta-analysis was conducted on clinical studies assessing the effects of psychostimulants on emotional dysregulation [78], showing that MPH was significantly more effective than placebo in reducing symptoms severity with small-to-medium effect size. Additional evidence included a low effect size for Atomoxetine and a small-to-medium one for Lisdexamfetamine, though the paucity of studies conducted with these drugs did not allow for definite conclusions. However, the three compounds were significantly effective in reducing emotional dysregulation symptoms in addition to the core symptoms of ADHD.

In our opinion, in most of the cases, a two-step treatment approach is advisable, with mood stabilization preceding the treatment of ADHD symptoms. Stimulants are the most effective medications for ADHD, but their use may be contraindicated in the presence of a comorbid drug abuse or in patients that simulate or exaggerate ADHD symptoms in order to obtain stimulants for diversion or abuse [66]. Evidence supports positive effects on suicide prevention in ADHD patients treated with MPH, particularly when mood disorders are associated. ATX can be used in patients who do not respond to stimulants, or when stimulants are contraindicated (i.e., worsening of tics in Tourette Syndrome, evidence of negative effects on growth and weight). Based on empirical evidence, the low use of ADHD medications in BD youths may have negative short- and long-term implications on outcome.

4. DISRUPTIVE BEHAVIOR DISORDERS

A bidirectional association between BD and DBD, including ODD and CD, has been consistently reported in clinical, epidemiological, and research samples [79]. Comorbidity rates according to data from a large, unselected sample of referred children and adolescents were estimated as high as 40% of CD youths with BD and 41% of BD youths with CD [80], Biederman and colleagues closely explored the phenomenology of youths with comorbid CD-BD using a structured interview, and found that the presence of one disorder did not affect the presentation of the other, except aggressiveness being higher in those children with both disorders [80]. Conversely, a more recent study showed that the comorbidity between BD and ODD/CD is featured by higher rates of impulsive aggression and higher risk of substance abuse [81]. Regarding the type of aggression in CD and BD, Vitiello and Stoff previously suggested that aggression is more frequently impulsive, non-predatory and disorganized in BD, and purposeful, predatory and organized in CD, although both subtypes can be present in both disorders particularly when they are associated [82]. Furthermore, BD youths with comorbid CD manifest poorer recovery from first manic or mixed episodes as well as greater number of mood episodes [83]. Further studies on this association may be useful, given the prevalent role of psychosocial variables in affecting this risk in different countries.

Interestingly, such comorbidity frequently overlaps with the concomitant occurrence of ADHD [84]. In a study on boys with ADHD, 55% of those with CD also had BD and 71% of the boys with BD also had CD [85]. Indeed, the rate of CD is significantly greater in children with comorbid BD and ADHD compared to children with ADHD alone [32], with comorbid ADHD-BD representing a major risk for the development of subsequent ODD/CD and substance abuse [42]. Vice versa, comorbid ADHD-CD is a risk factor for developing subsequent mood disorders [39]. Youths with comorbid BD, CD and ADHD were most severely impaired according to the number of hospitalizations and characterized by an acute and explosive type of aggression, mainly explained by the manic component, beyond the basic impulsive aggressiveness of CD; rates of alcohol and substance dependence are also higher [81].

Negative prognostic implications including a poorer response to treatments are frequently found in youths with comorbid CD-BD. Both naturalistic studies and open-label trials found that DBD comorbidity is a major predictor of poor treatment response for manic symptoms in monotherapy with mood stabilizers [86, 87]. However, a randomized controlled trial indicated that youths with comorbid DBD-BD experienced greater improvement in manic symptoms when treated with antipsychotics versus mood stabilizers, whereas youth with non-comorbid BD experienced similar symptoms improvement with both medications [88].

There is a shared consensus on the relative usefulness of medications in ODD/CD directly targeting core symptoms of the disorders, though prescription of certain psychotropic drugs, especially antipsychotics, increased considerably in the last decades, particularly second generation antipsychotics [89]. Lithium may be effective on aggression in the context of CD with a high level of evidence though provided by studies with limited sample sizes and short follow-up [90, 91]. Nonetheless, no medications are currently approved in the US to treat CD, while in Europe, the evidence supporting Risperidone treatment for CD and aggression has been judged sufficient to warrant its approval. Treatment with psychostimulants is also highly recommended for patients who have both comorbid ADHD and conduct problems.

While only few pharmacological treatments for CD are available, BD can be successfully treated, and its treatment may even mitigate symptoms of CD in children affected by both disorders [79]. Considering the high morbidity and disability associated with CD and the limited treatment options available, these findings would have very important clinical and public health significance. Nevertheless, the presence of comorbid CD is among the predictors of non-response to both Lithium and Valproic Acid [86, 87], while Quetiapine and Risperidone, instead, may be promising agents with potential use in young patients, with similar efficacy in reducing manic symptoms and aggression [92, 93]. Quetiapine was more effective in improving anxiety and depressive symptoms, including suicidality, while changes in body mass index (BMI) and prolactin levels were more evident with Risperidone [94]. Long-acting intramuscular injectable preparations of second-generation antipsychotics (SGAs) have also been tested in patients with BD and/or CD, who were non-compliant to oral antipsychotic drugs, with global clinical improvement both in comorbid and non-comorbid patients, but with weight gain, extra-pyramidal symptoms and galactorrhea frequently reported [95]. So far, controlled studies with larger sample size aimed at looking for long-term side effects and for the impact of treatment dropout are missing.

In summary, the co-occurrence of CD and other DBD is a negative predictor of response to medication, with more frequent need for antipsychotic agents, which are more effective but also show heavier side effects. However, improving manic or mixed symptoms can ameliorate the aggressiveness and impulsivity of bipolar patients, with positive effects on the short- and long-term outcome.

5. SUBSTANCE USE DISORDERS

The association between BD and SUD, including alcohol and drugs use disorders, has long been established in adult populations, both in clinical and epidemiological settings [96-98]. However, only in the last two decades, several studies on the mutual relationships between youth BD and SUD have emerged. Indeed, youth-onset BD was early recognized as a risk factor for SUD in retrospective studies showing higher SUD comorbidity rates in early-onset than in adult-onset BD [99, 100], and, among BD youths, in those with adolescent-onset compared to those with child-onset [101-103]. Subsequent cross-sectional clinical studies revealed a SUD point prevalence of approximately 20% among adolescents with BD [104], but up to 45-48% of BD adolescents were diagnosed with comorbid SUD at baseline and/or at follow-up in longitudinal studies [105, 106]. Indeed, prospective studies conducted on BD youths revealed that up to one third of non-comorbid patients are likely to develop a SUD during clinical follow-up: an incidence of SUD of 32% has been reported in a sample of children with BD type 1 followed for about 8 years [107]; similarly, 32% of adolescents with bipolar spectrum disorders, mostly recruited from outpatient clinics, and followed for on average 4 years, developed a first-onset SUD [105]. Impressively, almost one fourth of non-comorbid patients hospitalized for first-episode mania developed a SUD after approximately 40 weeks from hospital discharge [106]. In accordance with these findings, higher rates of comorbid SUD were repeatedly observed in BD youths compared to control subjects without mood disorders [103, 108, 109], and significantly higher rates of drug abuse were also found in adolescents with BD compared with those with Major Depressive Disorder, while a non-significantly higher rate of Alcohol Use Disorder (AUD) was encountered in one study [110].

Both drug abuse and alcoholism, together with psychosis, suicide attempts and social impairments, were retained as phenotypic features of BD showing familial aggregation among 40 candidates in a sample of 1246 adults from 172 multiplex that families [111]. Clinical studies confirmed that youth with BD show higher rates of SUD among family members compared to major depressive youths and non-affected controls, independently from SUD or CD in probands [104]. Moreover, the SUD-BD phenotype has been related to a higher familial load for BD in families of youth with BD type 1 in one study [112]. In the same sample, the SUD-BD type 1 phenotype has also been shown to run in families, since the comorbidity between mania and SUD was found to be overrepresented among first-degree relatives of SUD-BD type 1 adolescents [112]. Nevertheless, a recent study on adolescent and young adult relatives of patients with BD, showed that the increased risk of SUD associated with BD family history was no longer significant when accounting for youth psychopathology, particularly mood and externalizing disorders [113]. Conversely, in another study, additive effects of parental SUD and offspring BD on the risk of alcohol use disorder were observed [114].

Among adolescents with BD, several features, other than family history of SUD, may help identify those patients at higher risk of developing a SUD. A later onset of BD [101, 106], low family cohesiveness [105, 115], comorbidity with DBD [106, 115], particularly ODD [101, 105], and anxiety disorders [101], particularly panic disorder [105], were all repeatedly found predictive of SUD onset in longitudinal studies, while a greater (hypo)manic symptom severity in the 12 weeks preceding SUD onset was reported in one study [105]. Comorbid Post-Traumatic Stress Disorder (PTSD) was also significantly associated with developing a SUD in a prospective evaluation of adolescents with first-episode mania [106], while impactful child life events have been associated with regular drug use at follow-up in a sample of children enriched for elevated manic symptoms [115]. In addition, in a previous retrospective account, significantly more SUD was reported in BD youth with comorbid PTSD than in those without or with subthreshold PTSD [116]. In five comorbid cases PTSD preceded SUD, while in the remaining two cases the coincident onset of BD and SUD led to trauma and subsequent PTSD [116]. Two additional BD features showed a bidirectional longitudinal relationship with SUD. First, psychotic symptoms were among the strongest predictors of SUD development in a study on patients with first-episode mania [106]. Conversely, in a sample of BD offspring, psychotic symptoms were predicted by first-onset SUD, mostly Cannabis Use Disorder, independently from the diagnosis of BD [117]. Although most research on the association between cannabis and psychosis has been conducted in patients with schizophrenia spectrum disorders, a positive relationship between cannabis use and the presence of psychotic symptoms in BD patients was repeatedly observed in adult studies [118]. Secondly, a history of suicide attempts has been found to strongly predict the onset of SUD in a retrospective longitudinal evaluation of patients with youth-onset BD [101]. On the other hand, several cross-sectional adult studies observed a positive association between SUD and a history of suicide attempts [119], especially in patients with AUD [120-122] or cannabis use [118, 123], while a case-control study including 96 adolescent suicide attempters found other illicit drug use disorders (i.e., inhalants, cocaine and hallucinogens) to be significantly predictive of suicide attempts [124].

While several studies reported on the onset of SUD in patients with youth BD, Duffy and colleagues suggested that SUD may also precede major depressive or full-blown (hypo)manic episodes in those with genetic risk for BD [117]. Indeed, minor mood disorders (i.e., cyclothymia, dysthymia, adjustment disorder with depressive symptoms, depression or mood disorders not otherwise specified) and non-affective psychopathology prospectively predicted SUD occurrence in BD offspring, while a mutual longitudinal relations between full-blown BD and SUD were observed [117]. A peak hazard of developing SUD was found between the ages of 14 and 20, especially in subjects with BD or minor mood disorders [117]. These findings are in line with a previous representative community study of 1709 adolescents, showing a similar prevalence of SUD both in subjects with full-blown BD (N = 18) and in those with subsyndromal manic symptoms (N = 97) [125]. From a longitudinal perspective, subsyndromal manic symptoms have been found to significantly increase the risk for later onset of alcohol abuse/dependence, cannabis abuse/dependence, and benzodiazepine abuse/dependence in a representative sample of young adults [126], and, more recently, a high and rising 24-month manic symptoms trajectory significantly predicted regular alcohol use, though not regular drug use, over a 6-year follow-up of children recruited in the Longitudinal Assessment of Manic Symptoms (LAMS) study [115]. Similarly, childhood severe affective/behavioral dysregulation has been associated with an increased risk for SUD in young adults [127].

Unfortunately, effective treatments for SUD-BD comorbidity are still an area of unmet needs, since comorbid conditions have been systematically excluded from BD and, somewhat less, from SUD trials, especially in youth populations [104, 128]. Indeed, only one 6-week, double-blind placebo-controlled pilot trial has been conducted in adolescents with SUD-BD [129]. The study compared Lithium to placebo in a small group of patients with bipolar spectrum disorders and secondary substance dependence, mainly alcohol and cannabis [129]. Lithium monotherapy at a therapeutic blood level of approximately 0.9 mEq/L was well tolerated and was associated with improved global clinical impression and with more negative urine drug tests compared to placebo [129].

More randomized-controlled studies are available in the adult population, especially in patients with AUD-BD. However, only Valproic Acid preparations, as an add-on to Lithium treatment, repeatedly improved alcohol use outcomes, including a greater reduction of drinking days, a lower rate of relapse to heavy drinking, and lower gamma glutamyl transpeptidase compared to placebo [128, 130]. Conversely, Quetiapine has been tested in three randomized-controlled trials but proved useless for alcohol-related outcomes [128], though being effective on depressive symptoms in a sample of outpatients with AUD-BD [131]. A similar finding was recently reported for Ondansetron in a sample of adults with bipolar spectrum disorders and comorbid early-onset AUD [132]. Aripiprazole, instead, showed some superiority against placebo on secondary outcomes, including marginal effects on the number of drinks per drinking day and a time-limited decrease in percent carbohydrate-deficient transferrin, possibly suggesting that low doses may be preferred [133]. Among anti-craving medications, Naltrexone was effective in decreasing drinking days and alcohol cravings in AUD-BD patients [134], while only limited benefits over placebo were provided by Acamprosate [135]. As for other substances, promising results are available for Citicoline in BD patients with cocaine dependence, in terms of negative urine tests and cognitive outcomes, though differences from placebo declined over time [136, 137]. Finally, a recent 1-week, double-blind, placebo-controlled, crossover, magnetic resonance imaging pilot study found that Gabapentin was well tolerated in 22 adults with comorbid CUD-BD and may be considered, based on the observed modulation of brain GABA/glutamate levels, as a candidate adjuvant medication for cannabis use warranting further investigation [138].

Since only limited treatment strategies are available for SUD-BD comorbidity, the treatment of BD before SUD comorbidity occurrence has been proposed as a secondary prevention strategy worth of study [104]. Based on results from naturalistic longitudinal studies, pharmacological treatment of BD and its comorbidities may help reduce the risk of subsequent SUD. Indeed, a higher number of medications prescribed significantly decreased the risk of regular drug use over a 6-year follow-up in a cohort of children enriched for symptoms of mania [115]. In another study, antidepressant treatment was negatively associated with SUD at follow-up in a sample of adolescents with bipolar spectrum disorder [105]. More importantly, in the same study, the absence of Lithium exposure in the 12 weeks preceding predicted SUD onset, increased the likelihood of comorbidity [105]. Finally, adolescents treated with psychostimulant treatment before their first manic episode were significantly less at risk to develop a SUD, independently from ADHD diagnosis [106]. Although these findings are likely to be subject to referral bias and should be interpreted with caution, it seems reasonable that prompt treatment of BD and related comorbidities may lower the risk of developing a SUD, at least in some patients prone to self-medication [139].

In summary, as previously proposed by Goldstein and Bukstein, BD should be considered a modifiable risk factor for SUD, especially in the adolescent population [104]. Accordingly, all BD youths should be carefully and repeatedly screened for substance use. When multiple risk factors apply, mood stabilization should be improved. Importantly, in high-risk patients (e.g., SUD-BD family history, DBD and/or anxiety disorders, substance use), a mood-stabilizing treatment may be considered right after the onset of minor mood symptoms, before the first occurrence of full-blown affective episodes. Moreover, SUD can also be considered a prodrome of BD in those with genetic risk and may thus warrant mood stabilization.

6. AUTISM SPECTRUM DISORDER

It is increasingly recognized that comorbid psychiatric disorders may significantly contribute to the morbidity associated with ASD [140]. Among mood disorder comorbidities in ASD, BD has been found to represent the vast majority of cases, up to 75% [20, 141]. The overall prevalence of comorbid BD in ASD clinical cohorts has been estimated at 4-7% with no significant gender-related difference [142], as also confirmed by a large national registry study [143], while higher rates have been reported in adult patients [144]. Conversely, ASD prevalence in BD is estimated to range from 1.4% in population-based studies up to 30% in clinical samples [140, 145]. High-functioning ASD seems more closely related to BD and, among the relatives of these patients, a high prevalence of mood disorders has been found [144].

Moreover, even without a clearcut diagnosis of BD, the prevalence of subclinical manic symptoms, including elation, expansive mood, and racing thoughts, is rather high in ASD youths [146]. Notably, some authors consider BD to be often underestimated in ASD samples, especially because of the communication difficulties and low intellectual functioning that complicate the identification of comorbid clinical pictures [147]. To further increase the complexity of the issue, BD in ASD is often characterized by atypical characteristics, such as irritability and behavioral disturbances, which are frequently associated with several other potentially comorbid conditions, including, above all, ADHD [147].

Overall severity of mood episodes in youths with ASD-BD is similar to that of youths with non-comorbid BD, though an increased functional impairment related to comorbid ASD has been reported [145, 147]. A highly episodic course characterized by manic episodes with mixed features including irritable, unstable, and dysphoric mood, and subsyndromal mixed symptoms during inter-episode intervals, are frequently observed, while depressive episodes may show milder features and chronic course [144, 148]. During mania, grandiosity, hallucinations, psychotic interpretations and delusional ideas are dominant features of high-functioning ASD, while aggressive behavior, irritability, and overactivity are common in low-functioning ASD [140, 144]. Indeed, symptoms of aggression in ASD children are more likely to correlate with limited verbal skills and intellectual disability [149]. Overall, early onset of BD, mixed presentations, and greater functional impairments are evident in youths with ASD-BD [4].

BD comorbidity has also been linked with the occurrence of catatonia in subjects with ASD, since catatonia is often due to an exacerbation of mood symptoms, being in some patients the easiest clinical signs to detect mood disturbances [150]. Lifetime prevalence estimates of catatonia in hospitalized psychiatric populations range from 0.6% for adolescents to between 6-10% for adults [150], but in clinical practice, it is common to misinterpret catatonic symptoms as features of ASD due to the partial overlap of core symptoms. Nevertheless, a high prevalence of catatonic symptoms is reported among ASD individuals, up to 17%, when catatonia is considered based on changes in functioning and psychomotor activity [150, 151]. Patients with autism and catatonia preferentially show features such as echolalia and echopraxia, stereotypies and mannerisms, mutism, negativism, and repetitive psychomotor behaviors [151]. A difficult differential diagnosis between catatonia and Neuroleptic Malignant Syndrome (NMS) after antipsychotic treatment may occur in ASD patients, and should be always considered in this specific population, considering the effectiveness of catatonia treatments on the NMS symptoms [152]. The intrinsic difficulties of exploring NMS in ASD in controlled studies account for the subsequent lack of available information.

Finally, suicidal behaviors are also frequently observed in ASD-BD patients, and may confer a greater risk of hospitalization concerning non-comorbid subjects [149, 153]. Notably, ASD is an independent risk factor for suicide [154] with a probability to die by suicide which is 10 times greater than in the general population, both due to the tendency to a higher vulnerability to stressors, possibly related to impairments in interpreting social relationships, and to a high lethality and violence of suicidal methods used by individuals with ASD [155]. This latter evidence may be at least in part explained by a less impulsive nature of suicidality in ASD-BD compared to BD, which implies a colder transition from ideation to a suicide attempt and subsequent more lethal planned strategies [154, 155]. Suicidality is especially frequent in those with high-functioning ASD as high cognitive competencies may drive an increased self-consciousness of one’s disability leading to feelings of inadequacy, guilt, and exclusion by peers [155].

The pharmacological treatment options available for ASD are not directed to core symptoms of the disorder but rather target challenging behaviors associated with comorbidities [156]. Second-generation antipsychotics Risperidone and Aripiprazole are currently approved by Food and Drug Administration in the USA and European Medicine Agency in Europe in treating irritability and aggression in ASD youths, with no significant difference in effectiveness and tolerability [157-159]. These agents are also commonly used in youth with BD, and are particularly indicated in comorbid ASD-BD, namely when impulsive aggression is prominent [159]. Despite an increased severity of BD symptoms in ASD youths, the presence of this comorbidity does not impact the relative rate of response to irritability. Indeed, approximately 65% of ASD-BD patients treated with SGAs showed a greater than 30% improvement in irritability, which is similar to the non-comorbid BD peers [145]. Given the high risk of side effects, firstly weight gain and metabolic disorders, increased serum prolactin and sedation, antipsychotics should be used cautiously [160].

Although controlled trials are substantially lacking, numerous case reports and series support the efficacy of Lithium for mood instability, impulsivity and cyclicity in ASD subjects [144]. Lithium is often recommended as the first choice for the treatment of manic symptoms, especially in those with positive family history of BD. Lithium seems to be effective in controlling aggression and overactivity, as well as in reducing manic symptoms and mood swings. In a recent retrospective chart review, 73.7% of patients with ASD experienced significant clinical improvement after add-on Lithium salts to their treatment regimen, thus representing a reliable alternative to antipsychotics and other psychotropic medications [161]. Valproic Acid is the anticonvulsant with the largest evidence in comorbid ASD-BD, with particular efficacy in alleviating hyperarousal symptoms, irritability, dysphoria and anxiety [162]. As a conclusion, mood stabilizers may be thus preferable to antipsychotics, especially when aggression and hostility are not prominent, and antipsychotics should be used at the lowest effective dose and for strictly necessary periods to minimize adverse effects.

7. ANXIETY DISORDERS

The comorbidity between BD and ADHD or DBD has received greater attention in children and adolescents, comorbid anxiety in pediatric BD is a neglected but frequent and clinically relevant phenomenon. According to a recent meta-analysis, 44.7% of patients with pediatric BD show at least one lifetime comorbid anxiety disorder, with Generalized Anxiety Disorder (GAD) being observed in 27.4% of patients, Separation Anxiety Disorder (SeAD) in 26.1%, Social Anxiety Disorder (SoAD) (formerly known as Social Phobia) in 20.1%, and Panic Disorder (PD) in 12.7% [3]. Since early studies, Multiple Anxiety Disorders (MAD) have also been often observed in pediatric BD samples [8, 163, 164]. Consistently, among 446 youths with BD recruited as participants for the COBY study, 18% met the criteria for MAD [17]. A similar prevalence of MAD (22%) has also been confirmed in a large epidemiological cohort of 8,129 youths with BD [165].

The high prevalence of anxiety disorders in pediatric BD is consistent with the finding that the offspring of adult bipolar probands often initially receive anxiety disorder diagnoses [163, 166]. Anxiety disorders may represent a correlate or an early precursor of pediatric BD as suggested by longitudinal prospective studies [167, 168]. Indeed, based on recent meta-analyses, the comorbidity of any anxiety disorder, SoAD, GAD and SeAD, with pediatric BD, is higher than the comorbidity found in adult cases, while only PD shows higher rates in adults [3, 169]. Similarly, the offspring of parents with BD usually show SoAD, GAD and SeAD in early or middle childhood, while PD and BD are more frequently observed in adolescence [170].

Significant differences between SoAD and PD have been observed when examining the temporal relationship between (hypo)mania and anxiety comorbidity in adult patients [171]. On one hand, the onset of SoAD is almost invariably preceded by years by the onset of BD and never overlapped with the first (hypo)manic episode. Complete remission of SoAD was found during (hypo)manic episodes in almost all cases [171], consistently with the common occurrence of (hypo)mania in patients remitting from SoAD after antidepressant treatment [172]. Consistently, patients with SoAD with good response to antidepressants have been found not to differ in terms of clinical features from patients diagnosed with BD type 2 during mild/moderate hypomanic episodes [173]. Therefore, it is possible to hypothesize that SoAD may represent itself, at least in a subgroup of patients, the precursor of an atypical bipolar spectrum condition, where depressive inhibition is manifested by social inhibition and anxiety, and hypomanic episodes are usually drug-induced [172-174].

On the other hand, instead, PD has been found to precede BD in 27-33.3% of cases [171, 175, 176], while overlapping with the first (hypo)manic episode in 28.6% of cases [171]. Moreover, PD has been associated with short hypomanic episodes in adults with recurrent brief depressions [177] and with a higher number of hypomanic episodes in patients with BD type 2 [178]. Anxiety sensitivity, a known risk factor for PD which consists of excessive fear of anxiety sensations and beliefs that these sensations are harmful, may explain the interaction between BD and PD, since a state-dependent increase in anxiety sensitivity has been observed during (hypo)manic and mixed states [175].

Besides theoretical issues, comorbid anxiety disorders show relevant clinical and prognostic implications in patients with BD. Indeed, they have been associated with female gender, younger age at onset, higher depression severity, somatic complaints during depression, poorer global functioning and quality of life, shorter euthymia, slower recovery from depression, and higher risk of early relapse both in adults [178-182] and in children and adolescents with BD [8, 17, 165, 176, 183-186]. Several studies have also repeatedly highlighted the association between comorbid anxiety or anxious symptoms and suicidal ideation, suicidal behavior, and past suicide attempts in youths with BD [110, 184, 185]. Notably, this association was still significant when controlling for demographic confounders and psychiatric comorbidities in a recent study conducted on 9,720 adolescent inpatients with bipolar depression [185].

Despite the clinical importance of BD-anxiety comorbidity, no randomized-controlled studies have been conducted so far to assess the efficacy of pharmacological treatments of anxiety disorders in children and adolescents with BD. The same issue applies to adult BD, in which, to the best of our knowledge, only four randomized-controlled trials on the treatment of BD and comorbid anxiety disorders have been conducted. On one side, two 8-week, double-blind, placebo-controlled trials failed to support the anxiolytic effects of Risperidone and Ziprasidone monotherapy, respectively, in 111 and 49 adult patients with co-occurring PD or GAD [187, 188]. Conversely, significant improvements in anxiety were observed in a 12-week, single-blind trial of 47 Lithium-treated euthymic patients with current anxiety disorders randomized to add-on treatment with Olanzapine or Lamotrigine [189]. Both medications significantly reduced anxiety, but Olanzapine exceeded Lamotrigine in specific analyses. Finally, patients with BD and comorbid PD or GAD treated with Quetiapine extended-release were found to achieve greater improvements in anxiety than patients treated with Valproic Acid or placebo in an 8-week, double-blind, placebo-controlled trial including 149 patients [190].

Based on these findings, Olanzapine and Quetiapine may have some beneficial short-term effects on anxiety symptoms in adult BD patients, when compared with antiepileptic mood stabilizers (Lamotrigine and Valproic Acid), likely due to their sedative properties related to the blockade of a wide range of receptors, including histamine H1 receptors [191]. Indeed, several other randomized-controlled trials on the efficacy of atypical antipsychotics for anxiety symptoms in adults with BD have been conducted, with multiple studies showing modest efficacy of Quetiapine over placebo [192]. However, only a few studies are available overall, and most are limited by using outcome measures focusing on any type of anxiety morbidity, while none was directed at specific anxiety syndromes.

Further information on treatment response in children and adolescents with BD and comorbid anxiety disorders comes from longitudinal observational studies. In naturalistic settings, children and adolescents with BD and comorbid anxiety disorders have shown higher rates of treatment with antidepressants than BD patients without anxiety [8, 186]. In a longitudinal study, the persistence of anxiety disorders in the long-term was weakly though significantly associated with less follow-up time receiving antidepressant medications [193]. While these results may support the use of antidepressants in adolescents with BD and comorbid anxiety, a specific increase in treatment-emergent (hypo)mania in comorbid patients has been repeatedly reported. In an early study on 43 adolescent outpatients with BD, 36% of subjects with comorbid anxiety disorders, but none of those without, reported a history of treatment-emergent (hypo)mania [164]. A subsequent study confirmed that early-onset anxiety disorders were predictive of treatment-emergent (hypo)mania in pediatric BD [194]. The frequent use of antidepressants in comorbid patients could also be mediated by the association between anxiety comorbidity and BD type 2 [8,176], given that patients with BD type 2 spend more time with syndromal or subsyndromal depressive symptoms [1].

These findings suggest the need to be cautious when considering antidepressant pharmacotherapy in youths with anxiety disorders and BD. Close monitoring of possible worsening after antidepressant monotherapy, such as irritability, behavioral dyscontrol, substance abuse, and suicidality, is warranted. Whenever possible, a more conservative approach may be preferred, including considering the anti-anxiety properties of antiepileptic mood stabilizers. Based on adult studies, Gabapentin may have beneficial effects in SoAD and PD [195, 196], while Valproate may be effective in PD and anxious symptoms of BD [197]. Divalproex Sodium has also proved useful in reducing anxiety in a small sample of adolescent inpatients with BD [198], and, in a sample of 158 (hypo)manic youths initially treated with Valproate monotherapy, comorbidity with GAD, SeAD, and Simple Phobias were associated with a lower rate of switch to polypharmacy, while GAD was also associated with better response [87].

Considering the high risk of side effects after antipsychotic treatment, mood stabilizers with higher anti-anxiety properties (Valproate, Gabapentin) may be first considered. In treatment-resistant patients, Quetiapine may be considered in monotherapy or as an add-on treatment. Antidepressants should be avoided in monotherapy, and cautiously prescribed as an add-on treatment with stabilized mood and persisting anxiety, with discontinuation of a few weeks or months after the improvement of the anxiety symptomatology.

8. OBSESSIVE-COMPULSIVE DISORDER

The comorbidity between BD and OCD has been early proposed as a clinically relevant condition in adult patients. Since the beginning of the 20th century, the classical authors in French psychiatry debated the relationship between obsessive symptoms and affective syndromes, especially emphasizing the prevalence of obsessive phenomena in the context of melancholic and mixed depressive phases and subsuming episodic obsessive syndromes under the manic-depressive domain [199, 200]. More recently, Chen and Dilsaver first highlighted a specific link between OCD and BD [201]. Based on the Epidemiological Catchment Area database, they reported that 21% of adult patients with BD presented lifetime comorbidity with OCD. In addition, the odds ratios of subjects with BD meeting the criteria for OCD relative to subjects with unipolar depression or with any other major disorder were, respectively, 2.0 and 3.2 [201]. A meta-analysis of 46 studies confirmed similar estimates of comorbidity, showing a 17% pooled prevalence of OCD in BD and a similar rate of BD in OCD (18.4%) [202]. Notably, a lower mean age predicted a higher prevalence of OCD in BD patients and higher OCD prevalence rates were found in children and adolescents (24.2%) compared to adult patients (13.5%) [202]. According to a more recent meta-analysis, 16.7% of patients with pediatric BD show OCD comorbidity [3]. Even in this case, the comorbidity of OCD was higher than that observed in adult cases (9.7%) by the same authors [169] and higher comorbidity rates were reported by adolescent-onset studies than by childhood-onset studies [3].

While mutual interactions between OCD and BD have been highlighted in adult subjects, fewer studies were conducted in youth samples. Based on adult data, comorbid BD in adults with OCD has been associated with an episodic course of OCD, with obsessions usually worsening during mixed and depressive episodes and improving during (hypo)mania; higher rates of aggressive, sexual and religious obsessions, forbidden thoughts, poorer insight, greater comorbidity with body dysmorphic disorder, tic disorders and panic disorder, less favorable response to antidepressants and more frequent need for multiple pharmacological interventions were also observed [19, 203-212]. Compulsions of hoarding have been also repeatedly linked to BD comorbidity in adults [19, 206] and, more consistently, in children and adolescents [5, 213-215]. On the other hand, OCD comorbidity has been associated with a more chronic course of BD, a greater comorbidity with panic and other anxiety disorders, impulse control disorders and tic disorders, a higher number of depressive episodes and a history of treatment-emergent mania in adults with BD [204, 216-218]. Overall, comorbid OCD-BD patients show more frequent hospitalizations, greater impairments of functioning and lower quality of life compared non-comorbid counterparts both in adult [204] and youth samples [213, 214, 219].

While adult studies are not conclusive with respect to age at onset [204], more evidence comes from studies on children and adolescents and longitudinal investigations. Early studies from our group indicated that OCD in children and adolescents with co-occurring BD was associated with an earlier onset compared to non-comorbid OCD, while the age of onset of BD was not affected by comorbidity [213, 219]. Accordingly, OCD symptoms usually antedated the onset of BD in comorbid patients [213, 219]. A more recent study, instead, found no differences in age at onset of both BD or OCD between comorbid and non-comorbid subjects [5]. Interestingly, these discrepancies may be explained by the relationship between the primary disorder of ascertainment and earlier onset suggested by Joshi and colleagues [214]. According to their study, in comorbid patients, the disorder that emerged first was associated with worse impairment. As a consequence, those primarily referred for BD more frequently showed an earlier onset of BD, while among those primarily ascertained for OCD, BD had a later onset and followed OCD.

Nevertheless, a substantial progression from OCD to BD has been recently suggested based on two large longitudinal studies. In a Swedish register-based study, indeed, subjects with OCD showed a 12-fold increased risk of receiving a later diagnosis of BD, with a median of 2.7 years between the two diagnoses [220]. Even when controlling for the use of selective serotonin reuptake inhibitors (SSRIs), the risk of BD was still increased by almost 9 times [220]. In addition, a longitudinal community study, including 3021 adolescents and young adults prospectively followed for up to 10 years, found prior OCD to be associated with a 7-fold increased risk of BD [221]. These findings are supported by previous studies indicating that the association of co-occurring MAD (formerly including OCD), usually preceding the onset of bipolarity, may represent a possible pathway to specific bipolar phenotypes [8, 164, 170]. Indeed, significantly higher rates of comorbidity with MAD have been reported in comorbid OCD-BD patients (65-74%) compared to non-comorbid BD and OCD youths (35-37%) [214].

At the same time, BD comorbidity in children and adolescents with OCD has been repeatedly associated with higher rates of ADHD and DBD [5, 213, 214, 219]. Similarly, comorbid patients with ADHD and OCD have shown significantly higher rates of BD comorbidity than non-comorbid OCD youths [215, 222, 223]. Importantly, both BD and DBD comorbidity, especially CD, have been associated with a poorer response to treatment, as well as with a higher need for polypharmacy, in patients with OCD [213, 224] and, among comorbid OCD-BD patients, conduct disorder has also been associated with a poorer response [5]. Notably, in one study, both ODD/CD, together with hoarding symptoms, have been found to predict treatment non-response in OCD patients, while BD comorbidity was found to be irrelevant [213]. The negative effect of BD on treatment response may, thus, be mediated, at least in a subgroup of patients, by the comorbidity with DBD.

Interestingly, a clinically distinct juvenile behavioral phenotype of BD characterized by obsessive fear about harm to self or others, as well as excessive anxiety, oppositional behavior, aggressive tendencies, and frequent suicide threats, has been proposed by Papolos and colleagues [225], and has been associated with an early age of onset, severe manic and depressive symptoms, early and frequent psychiatric hospitalizations, significant social impairment and school problems [226]. According to these authors, a specific developmental sequence of fear-based behaviors could be clinically delineated in these patients, including night sweats, recurrent night-terrors and vivid nightmares, obsessive bedtime rituals, nyctophobia, separation anxiety, hypervigilance, misperception of neutral stimuli as threatening, reactive aggression in response to limit-setting or perceived threat or loss. When treatment resistance to mood-stabilizing agents and atypical antipsychotics is demonstrated, intranasal Ketamine every 3-4 days at sub-anesthetic doses may be a beneficial and well-tolerated treatment for these patients, according to a retrospective chart review of 45 patients followed for 3 months to 6.5 years [227].

Though augmentation with other medications acting on the glutamatergic system, such as Memantine and Topiramate, have proved some efficacy on obsessive and compulsive symptoms in adult manic patients recruited by two randomized placebo-controlled clinical trials [228, 229], most studies on the treatment of OCD-BD comorbidity focused on mood-stabilizing, antidepressant and antipsychotic agents [230, 231]. According to naturalistic studies, mood stabilizers are virtually required in all patients, both in adults [207] and youths [5]. Among adult patients, up to 42% of patients have been reported to require a combination of multiple mood stabilizers, usually Lithium plus antiepileptics [207]. Accordingly, comorbid youths are more frequently prescribed with mood stabilizers and less frequently with antidepressant monotherapy with respect to non-comorbid OCD patients [213]. Indeed, BD comorbidity in adults with OCD has been associated with high rates of drug-induced (hypo) manic switch under antidepressants (39-61%) [207, 232], especially in subjects not receiving mood stabilizers, with greater risks posed by Clomipramine compared SSRIs [207]. Despite this, it is still not clear whether OCD comorbidity in BD patients directly contributes to an increase in the risk of destabilization or whether only an increased exposure to antidepressants is responsible for such high rates of (hypo)manic switch.

As previously hinted, OCD-BD comorbidity in children and adolescents has been related to non-response to usual treatment, both compared to non-comorbid BD and OCD counterparts [5, 233], and SGAs are required in more than half of patients to manage comorbidity, suggesting a specific severity and treatment refractoriness [5]. Importantly, in two open-label trials, a greater efficacy of a combination of serotonin reuptake inhibitors with SGAs, namely Olanzapine and Risperidone, was observed in adult patients affected by treatment-resistant OCD with comorbid BD, respectively compared to non-comorbid treatment-resistant OCD [234] and subjects affected by treatment-resistant OCD and comorbid major depression [235]. Moreover, two randomized controlled trials on OCD-BD comorbid adult patients, demonstrated that Aripiprazole and Quetiapine augmentation widely exceeded placebo in treating obsessive and compulsive symptoms during manic episodes [228, 236].

Based on these findings, mood stabilization should be the primary focus when treating OCD-BD comorbidity. When mood stabilizing treatment has been titrated, OCD-specific treatments (i.e., serotonin reuptake inhibitors) may be employed, preferably at the lowest effective dose and carefully monitoring for (hypo)manic symptoms and affective instability. In case of non-response or reduced tolerability (e.g., drug-induced switch), or whenever patients show unfavorable prognostic features, such as DBD comorbidity and/or compulsions of hoarding, SGAs, such as Aripiprazole, Risperidone and Quetiapine, may be preferred, respectively as third- or second-choice options. Finally, a limited subgroup of comorbid patients may require a combination of multiple treatments, including mood stabilizers, antidepressants, and SGAs.

9. EATING DISORDERS

A growing literature shows that ED are common in BD patients, and individuals with BD have an increased risk of developing ED compared to the general population [237]. Lifetime prevalence of ED in the general population is approximately 3.8%, with lower rates (0.3-0.6%) for Anorexia Nervosa (AN), intermediate rates (1.5%) for Bulimia Nervosa (BN), and slightly greater rates (2.3-2.8%) for Binge Eating Disorder (BED) [238, 239]. A higher prevalence of ED was found by a recent meta-analysis in patients with BD: AN was reported in 3.8% of patients, BN in 7.4%, and BED occurred in 12.5% [22]. On the other hand, a diagnosis of BD was reported in 2% of patients with AN, 6.7% of subjects with BN, and 9.1% of patients with BED. Moreover, combined results from five studies comparing ED rates in BD type 1 versus type 2 suggested that ED rates are similar across BD diagnoses [22]. Of note, in a study considering DSM-5 criteria for ED and BD in a consecutive sample of 1,092 bipolar patients, higher rates of comorbidity were reported, as 27% of the sample satisfied criteria for current ED, 12% for BED, 15% for BN, and 0.2% for AN [240].

According to adult studies, ED with binging/purging symptoms, such as BED, BN and binging/purging AN, show the highest prevalence in patients with BD [241, 242]. In adolescents, subthreshold symptoms of binge eating or binging/purging are often reported, and, similarly, subsyndromal mood symptoms are frequently associated with ED [243], underscoring the need to examine the full spectra of BD and ED when assessing the co-occurrence of these syndromes. Indeed, as previously observed for anxiety disorders and SUD, ED may precede the onset of BD in up to 41.4% of the ED-BD cases [244], particularly among females [245]. Importantly, depression and (hypo)mania are often associated with changes in appetite and weight, partially mimicking the fluctuations seen among people with ED. However, these fluctuations should be distinguished from true comorbidity between BD and ED. BD patients with comorbid ED are more frequently females, younger, with an earlier onset of BD, and present with more anxiety, suicidality and mood instability, higher BMI, more frequent obesity, and higher general medical burden [240]. The comorbidity is associated with worse illness course and outcome, and an increased risk of additional clinical comorbidities, including alcohol and substance use disorders [246, 247].

The etiological mechanisms of the overlap are relatively unknown and possibly include common genetic factors [248], as well as exposure to certain psychotropic medication inducing ED [249]. Importantly, trait impulsivity and emotional dysregulation may be considered shared underlying traits stemming from a common pathophysiological basis, influencing the symptom expression of both disorders [246], closely related to the development of binging/purging behaviors [241], but also to comorbid alcohol/substance use disorder and suicidality [22, 246]. In this respect, BD and ED may be viewed as part of the same neurodevelopmental process affecting emotional regulation and impulse control abilities [22]. Further research is warranted to evaluate the role of emotional dysregulation as a precursor of ED-BD comorbidity and as a mediator of the relationship between each condition across the lifespan. Moreover, differences in emotional dysregulation may help explain differences between BD patients with and without comorbid ED, including those observed in gender, personality disorders comorbidity, maladaptive traits, rapid-cycling course, and suicidality, and may have relevant implications for future interventions [22, 245].

Patients presenting with ED-BD comorbidity require specific treatment considerations [241]. Failure to recognize and treat BD comorbidity in patients with ED may negatively affect the management of their conditions, as the undertreated BD symptomatology, including affective lability and impulsivity, can worsen ED symptoms and decrease adherence to treatments. The stabilization of affective symptoms may have a beneficial impact on the management of ED, improving the compliance with non-pharmacological treatments (i.e., familial-based interventions, psychotherapy, dietary interventions). While available evidence from RCT supports the efficacy of pharmacological treatments in adolescents with BD [250], the evidence of the efficacy of drug treatments for ED has been mostly weak [251]. Hierarchy and treatment strategy should be based on symptom severity, although often ED may be hardly effectively managed without improving clinically relevant depressive or manic symptoms.

Starting from these basic considerations, treatment strategies should consider approaches effective for each syndrome without adversely impacting the other, by avoiding treatments having iatrogenic effects. SGAs are increasingly used in adolescents with BD [252, 253], despite significant concerns regarding their metabolic side effects [250]. These compounds are also often used in AN, although empirical evidence is scarce [254, 255]. Moreover, in patients with BN or BED, SGAs, can worsen binge eating behaviors and favor weight gain. Antipsychotic medications may, indeed, disrupt signaling of neurotransmitters (e.g., serotonin) and hormones (e.g., leptin and ghrelin) associated with the regulation of hunger, raising appetite and increasing the risk of weight gain [256, 257]. Weight gain, by itself, may worsen eating behaviors both in BN/BED and AN. Both BN and BED are often associated with obesity, diabetes mellitus, and related metabolic disorders, hampering the management of comorbid BD with SGAs [258]. Finally, the significant cardiovascular risk associated with antipsychotic treatment should be carefully considered when these medications are administered to patients with ED and deteriorated physical conditions [259].

On the other hand, antidepressants, particularly Fluoxetine, have been widely used in the treatment of BN and BED in randomized, placebo-controlled trials [260, 261], while low evidence supports their efficacy in AN [262]. In these patients, who frequently present with associated osteoporosis, the increase in fracture risk observed in youths treated with antidepressants should be taken into account [263]. Moreover, these medications should be considered very cautiously in patients with BD, given the risk of emergent (hypo)manic switch or acute worsening of manic symptoms [264], even when associated with mood stabilizers [265].

Failure to recognize hypomanic symptoms in patients with ED, especially in BD type 2 patients with comorbid BN, may result in the prescription of antidepressant monotherapy, exposing to a further increase of affective lability, excitement and impulse dyscontrol, often co-occurring in BN and BED [266]. As for anxiety disorders, differentiating BD from major depressive disorder in adolescents is critical, since the recognition of previous (hypo)manic episodes may prevent the risks associated with the use of antidepressants (e.g., increased mood instability, impulsivity, and suicidality).

Unfortunately, mood stabilizers and anticonvulsants have not had a large role in the treatment of ED, and only limited evidence is available, mostly from adult studies. Lithium resulted effective in a small placebo-controlled trial in the treatment of AN [267], while both Lithium [268] and Carbamazepine [269] failed to exceed placebo in BN. However, add-on Lithium to Topiramate resulted in improvement of affective symptoms, weight loss and binge frequency in BED patients [270]. Topiramate was found effective in the treatment of BN [271] and BED [272], but its efficacy in BD is still poorly supported by evidence, both in adolescents [273], and in adults [274]. In addition, treatment-emergent (hypo)manic switches have been reported in adult patients treated with Topiramate [275]. Finally, Zonisamide, an anticonvulsant agent, showed positive effects on BED and obesity [276, 277], though its efficacy in BD has not been supported [278].

CONCLUSION

BD is a highly comorbid condition, and rates of co-occurring disorders are even higher in youth. Comorbid disorders strongly affect clinical presentation, natural course, prognosis, and treatment. Associations between BD and comorbid disorders define specific conditions which are not simply a sum of different clinical pictures, but rather occur as distinct and complex combinations with specific developmental pathways over time. These conditions require close attention to pharmacological treatments, with positive and negative implications (Table 1). The positive implications are related to the possible improvement of the whole clinical picture by addressing the comorbid condition rather than BD itself. The negative implication is the possible worsening of BD following pharmacological treatment of comorbidities (e.g., manic switches after antidepressants, depressive episodes or emotional flattening after antipsychotic treatments). Specific studies addressing the pharmacological management of BD and comorbidities are still scarce, and information is particularly lacking in children and adolescents. For this reason, this review included also studies with adult patients. Unfortunately, treatment strategies are often based on personal experiences or expert opinions. Developmentally sensitive controlled clinical trials are thus warranted to improve the prognosis of these highly complex patients, requiring timely and finely personalized therapies.

Table 1.

Summary of review findings.

Comorbid Condition	Main Recommendations
ADHD	MPH is greatly effective in improving ADHD core symptoms and emotional dysregulation with non-significant adverse effects, low risk of (hypo)manic switch and possible beneficial effects on suicidality
	ATX is effective in improving ADHD core symptoms with ongoing mood stabilizers (e.g., with Valproic Acid), but has the potential but still debated risk of inducing (hypo)manic switches and suicidality
	A two-step treatment approach is recommended with mood stabilizers prescription preceding the treatment of ADHD symptoms to prevent mood destabilization, especially in BD type 1
	MPH should be considered as first-line treatment option, followed by ATX when MPH is ineffective or contraindicated; other stimulants can be also effective after mood stabilization
	ADHD medications may be contraindicated in youths with comorbid substance use disorder, especially when ADHD symptoms are simulated or exaggerated to obtain stimulants for diversion or abuse
DBD	No medications are currently approved in US for CD, while Risperidone is approved in Europe for aggression; evidence favors antipsychotics over mood stabilizers for treating mania
	Lithium may be effective on aggression in CD, and the successful treatment of BD can mitigate CD symptoms, though comorbid CD is a predictor of non-response to both Lithium and Valproic Acid
	SGAs show similar efficacy in treating mania and aggression, but Quetiapine is more effective on anxiety, depression, and suicidality, while metabolic changes are more evident with Risperidone
SUD	Lithium is the only treatment with positive results in a randomized-controlled study of adolescents; Valproic Acid add-on may be considered for AUD, based on adult trials
	Quetiapine and Aripiprazole showed limited benefits in adult AUD; Citicoline and Gabapentin are promising options, respectively, for cocaine dependence and CUD
	Prevention of SUD should be the focus of treatment in patients with multiple risk factors; in high-risk patients a mood stabilizer may be introduced before the onset of full-blown BD
ASD	Pharmacological treatment does not address ASD core symptoms but rather targets challenging behaviors; the presence of comorbidity does not impact the rate of response to irritability
	Lithium may be effective for mood swings, (hypo)manic symptoms, impulsivity, and irritability, while VPA is the anticonvulsant with the largest evidence in the comorbid condition
	Risperidone and Aripiprazole are currently approved for treating irritability and aggression and are particularly indicated when impulsive aggression is prominent
	Mood stabilizers may be preferable especially when aggression is limited; antipsychotics should be used at the lowest effective dose and for the strictly necessary periods to minimize adverse effects
Anxiety Disorders	The use of antidepressants is supported, though with associated specific increased risk of (hypo)mania, which requires closely monitoring of irritability and suicidality
	Mood stabilizers with anxiolytic properties should be first considered in youths (Valproic Acid, Gabapentin)
	In the short-term Quetiapine and Olanzapine may exhibit greater beneficial effects on anxiety compared to antiepileptics
OCD	Antidepressants should be given at the lowest effective dose with careful monitoring of (hypo)manic symptoms due to high rates of switches, especially with Clomipramine
	Mood stabilizers, even in combination (e.g., Lithium and antiepileptics), are required in virtually all patients; BD should be the primary focus of pharmacological treatment
	SGAs are required in case of complex comorbidity, high severity, and refractoriness in more than half of patients; Olanzapine, Risperidone, Aripiprazole and Quetiapine are available treatment options
	A combination of mood stabilizers, antidepressants and SGAs may be indicated for selected patients
ED	Stabilization of affective symptoms may have a beneficial impact on eating behaviors, but the choice of effective approaches for each syndrome should consider avoiding iatrogenic effects on the other
	Antidepressants (i.e., Fluoxetine) should be given with caution due to increased risk of (hypo)manic switch, affective instability, impulsivity, and suicidality, even when prescribed with mood stabilizers
	Among mood stabilizers, Lithium may be effective in AN; Topiramate and Zonisamide have shown some efficacy for treating BED, but their use in BD is controversial
	SGAs should be avoided for their disruptive effects on hunger regulation, increased weight gain and dysfunctional eating behaviors, along with negative metabolic and cardiovascular effects

LIST OF ABBREVIATIONS

ADHD

Attention-Deficit/Hyperactivity Disorder

AN

Anorexia Nervosa

ASD

Autism Spectrum Disorder

ATX

Atomoxetine

AUD

Alcohol Use Disorder

BD

Bipolar Disorder

BED

Binge Eating Disorder

BMI

Body mass index

BN

Bulimia Nervosa

CANMAT

Canadian Network for Mood and Anxiety Treatments

CD

Conduct Disorder

COBY

Course and Outcome of Bipolar Illness in Youth

DBD

Disruptive Behavior Disorders

DSM

Diagnostic and Statistical Manual of Mental Disorders

ED

Eating disorders

GAD

Generalized Anxiety Disorder

LAMS

Longitudinal Assessment of Manic Symptoms

MAS

Mixed Amphetamine Salts

MAD

Multiple Anxiety Disorders

MPH

Methylphenidate

NMS

Neuroleptic Malignant Syndrome

OCD

Obsessive-Compulsive Disorder

ODD

Oppositional Defiant Disorder

PD

Panic Disorder

PTSD

Post-Traumatic Stress Disorder

SeAD

Separation Anxiety Disorder

SGAs

Second-generation antipsychotics

SoAD

Social Anxiety Disorder

SSRIs

Selective serotonin reuptake inhibitors

SUD

Substance use disorders

CONSENT FOR PUBLICATION

Not applicable.

FUNDING

This research was funded by the Italian Ministry of Health: Ricerca Corrente (Project 2.10: Bipolar spectrum in youth, from cyclothymia to bipolar disorder: clinical and pharmacological implications; GM), and the “5 per 1000” voluntary contributions.

CONFLICT OF INTEREST

Dr. Masi has received research grants from Lundbeck and Humana, was in the advisory board for Angelini, and has been a speaker for Angelini, FB Health, Janssen, Lundbeck, and Otsuka. All the other authors do not have conflicts of interest to declare.