Bipolar disorder (BD) is a severe, often chronic condition with lifetime prevalence rates of up to 6.5% for bipolar spectrum disorders in the general population (1). BD patients frequently report co-occurring substance-use disorders (SUDs) and behavioral addictions (1–5). The rates of alcohol and other SUDs are significantly higher in subjects with BD than in the general population (1, 6). The co-occurrence of BD and addiction has important clinical implications (3, 7). Bipolar patients with comorbid conditions present with a more severe course of illness (8), characterized by an overall worse clinical picture (9), poorer treatment outcome (10–12), higher suicidality (13), and mortality (14).

Several studies have aimed to identify the endophenotypical features predisposing to the development of addiction in the general population, as well as in the context of BD. These studies focused on genetic vulnerability, impulsive traits, and decision-making impairment (15–19).

The aim of this paper is to present the possible contribution of the concept of allostasis as a framework linking BD and addiction. We hypothesize that the concepts of allostasis and allostatic load (AL) may contribute to the understanding of the complex relationships between BD and addictive behaviors (20–22). Allostasis entails the safeguarding of reward function stability by recruitment of changes in the reward and stress system neurocircuitry (21) and it may help to elucidate neurobiological underpinnings of vulnerability to addiction in BD patients.

Computerized database, i.e., PubMed, Psycinfo, Cochrane Library were searched using the following terms: “allostasis,” “AL,” “reward,” “hedonic tone,” “stress system” cross-referenced with “BD,” “addiction,” and “SUDs.” The results of this search are presented in this article, and examined in light of a unifying hypothesis with a potential heuristic value to inform and provide direction to future research in this intriguing area.

A complete review of neurobiological features in BD-SUD comorbidity is beyond the purpose of this paper. Familial and illness course characteristics of BD and addictive disorders, as well as shared underlying mechanisms suggest potentially important genetic overlap (19, 61, 62). Preliminary findings hint at the existence of a shared genetic vulnerability for BD and SUDs (15). Johnson et al. (63) found convergent genome-wide association results for BD and SUDs. Products of one group of these genes are likely to play substantial roles in the initial and/or plasticity-related “wiring” of the brain (63). A second group of genes is the family of clock genes, implicated in the regulation of behavioral and physiological periodicity (19). Recently, a significant genetic overlap between candidate genes for both alcoholism and BD was found (64, 65), by using the d-box binding protein knockout mouse, a stress-reactive animal model developed consistently with allostasis and stress-surfeit theory of addiction (46).

To date, no studies have specifically investigated neuroimaging correlates in comorbid BD–SUD patients. Several studies describe putative mechanisms involved in BD vulnerability to addiction. Structural imaging studies in BD patients found volume reductions in prefrontal cortex [PFC (66)], which is involved in encoding incentive information (67). During Iowa gambling task (IGT), BD patients showed abnormalities in the dorsal and ventral PFC, while lateral temporal and polar regions displayed increased activation (68). Jogia et al. (69) confirmed these observations and also reported a greater activation in the anterior cingulate cortex of BD patients performing the IGT and in the insula during the n-back working memory task. Reduced functioning of the dopamine transporter (DAT) has been linked to BD (70–72). Animal models may provide insight into the role of the dopaminergic system in risk-taking behavior. Mice with reduced DAT functioning exhibit a behavioral profile consistent with manic patients and increased risk-taking behavior during a mouse version of the IGT (70). Evidence from these animal model studies and translational human research in BD subjects (73, 74), allows us to hypothesize that system-related change involving functioning of the dopamine system play a role in impulsive choice, risk-taking behavior, and reward, thus help guiding future studies in BD–SUD subjects.

Dopaminergic mechanisms are likely to play a key role in the understanding of the pathophysiology of BD and the clinical phenomena of mania and depression have previously been conceptualized in terms of an increase or a decrease in dopaminergic function, respectively (75, 76). Also, converging lines of evidence suggest that dopamine is a key neurotransmitter mediating hedonic allostasis in drug and behavioral addictions (49, 77). From a neurobiological perspective, a central dopaminergic dysfunction has been widely proposed as a neurobiological correlate of anhedonia (78). Different studies suggest anhedonia as a key symptom in addictive disorders, both as part of a withdrawal syndrome and as a relevant factor involved in relapses (51, 59, 79). In addition to dopamine, other neurotransmitters are believed to encode the hedonic experience [endogenous opioids, serotonin (80)], while long-lasting alterations involving cue-induced craving and relapse are thought to result from neuroplastic changes in glutamatergic circuitry (81–83).

Several studies provide support for reward dysregulation accounts in BD (16, 18, 45, 69, 84–95) (Table 1), characterizing neural dynamics underlying inter-temporal reward processing (90). Possibly emotional dysregulation present in BD is related to hypersensitivity to reward-relevant stimuli (93). Impulsive and unsafe decision-making in BD is linked to decreased sensitivity to emotional contexts involving rewards or punishments, possibly reflecting altered appraisal of prospective gains and losses associated with certain behaviors (89). It has been proposed that anhedonia could be mediated by a change in reward sensitivity (78), which has different behavioral consequences involving either stress-related and dopaminergic processes (96). In BD, sustained allostatic states and the consequent cumulative brain damage resulting from increased AL may play a part in the occurrence of negative affective states (i.e., anhedonia) that persist even during periods of remission (84). Counter-adaptive processes, such as opponent process that are part of the normal homeostatic limitation of reward function (55) fail to return within the normal homeostatic range and are hypothesized to repeatedly drive the allostatic state [decreased dopamine and opioid peptide function, increased CRF activity (49)]. This allostatic state is hypothesized to be reflected in a chronic deviation of reward set point that is fueled, not only by dysregulation of reward circuits per se but also by recruitment of brain and hormonal stress responses.

Altered functioning of the HPA axis may hold clues to the nature of the motivational changes accompanying addiction and vulnerability to addiction (97). Pre-existing alterations in frontal–limbic interactions with the HPA may reflect addiction-proneness, as shown in studies of offspring of alcohol- and drug-abusing parents (98). Alterations in the CRF/HPA axis may exert effects on the corticostriatal-limbic motivational, learning, and adaptation systems that include mesolimbic dopamine, glutamate, and gamma-amino-butyric acid (GABA) pathways (97), representing the underlying pathophysiology associated with stress-related risk of addiction.

The effects of these allostatic changes in the mesocorticolimbic brain system and in CRF/HPA axis contribute to the underlying pathophysiology associated with stress-related risk of addiction in BD (99). In BD patients, we hypothesize that the hedonic response to an acute drug administration occurs on a pre-existing allostatic dysregulation of the dopamine and CRF system. BD-related allostatic alterations in reward and stress systems thereby constitute vulnerability factors to the development of addiction in subjects exposed to occasional drug use. The failure to self-regulate these systems, determined by the collective contribution of endogenous factors linked to BD and of exogenous substances, results in an AL leading to a facilitated transition to drug addiction.

Dysphoria triggers drug intake, accompanied by an intense activity of the dopaminergic system and followed by a compensatory decrease in the dopaminergic system and increase in the CRF system to re-establish the allostatic set point. Such negative affective states may render BD patients more vulnerable to drug addiction, favoring a very rapid transition from occasional, recreational drug use to compulsive, pathological, drug dependence. The resulting addictive behavior-related ALs, in turn, may contribute to illness progression (Figure 1).

Converging data from addiction and BD studies suggest that these disorders involve similar allostatic processes, and allostasis can contribute to unify these disorders under a unitary perspective. In this context, the concepts of allostasis and AL provide both a pathophysiological model for the understanding of BD-addiction comorbidity and a new perspective for the development of novel therapeutic strategies for the treatment of comorbid patients (100, 101).

Allostatic alterations in brain reward system could render BD patients more vulnerable to drug addiction, favoring a very rapid transition from occasional, recreational drug use to compulsive, pathological, and drug dependence. This framework allows us to explain the high comorbidity rate between these disorders (2), as well as its relevance in early-onset patients (8, 102). Furthermore, it enables us to identify the factors of vulnerability to addiction in inter-episode periods as well (i.e., sub-threshold reward-system dysfunctions) (84). A more accurate monitoring of comorbidity-risk (103), coupled with the inclusion of specific tools for the assessment of hedonic tone, may contribute to early intervention on addiction-vulnerability factors and to initiate primary prophylaxis for substance misuse in youth suffering from BD with high-risk for addiction (104–106).

Currently, accruing evidence suggests that mood alteration episodes increase the risk of substance use (107, 108). Patients with dual disorders are more likely to use substances to self-regulate perceived internal factors (109, 110). SUD comorbidity in BD patients was preceded by greater manic symptoms in the previous period (104), as well as the persistence of depressive symptoms was associated with higher craving and increased risk to develop substance dependence (104, 108). Moreover, in gambling disorder (GD) patients depressive symptoms predicted gambling urges and duration (111). Allostasis framework enables us to extend the self-medication theory (112) beyond the established clinical domains, increasing the understanding of the interactions between BD symptoms and substance use. For instance, euthymic bipolar patients are more likely to experience cognitive impairment (deficits in measures of executive functions, verbal learning, immediate and delayed verbal memory, abstraction, sustained attention) (113). Cannabis abuse seems to positively affect cognitive function in a BD sample (114), and it may represent an attempt to counterbalance these alterations, even though causing an increased risk of rapid cycling and an earlier onset of manic episodes (114, 115).

Practitioners should be particularly vigilant in monitoring for substance misuse early after the onset of mood disorders, as well as they should be aware of personality traits related to the risk of addiction, in particular antisocial and schizotypal personality disorder (11, 116). The existence of additional risk factors [i.e., ADHD (117)] for the development of a BD-SUD comorbidity is controversial (105, 118). Combined with a specific role of traumatic stress as independent vulnerability-factor (99, 119), these elements contribute to the build-up of a cumulative AL. Clinicians can therefore incorporate specific therapy approaches for dual disorders (120–122) to target adherence weaknesses (123) and to enhance the effects of existing treatments.

Given the notion that exposure to stress or drugs leads to enduring changes in gene expression or activation of transcription factors, determining long-term neuroadaptation of brain functions, a promising field of future research could involve the detection of valuable markers of AL (124). In fact, markers of AL could contribute to prevention strategy (105, 116, 125); moreover, they could improve clinical monitoring and prognostic assessment of comorbid patients.

The clinical management of BD-SUD subjects requires a careful distinction between mood and withdrawal/intoxication symptoms (126, 127). Neuroimaging studies indicate that brain regions involved in mood regulation lie in close proximity to regions involved in motivation and craving (128). The complex interplay between addiction and BD domains, mediated by the involvement of similar neurobiological systems, requires further studies to better delineate how BD and SUD operate as reciprocal risk factors (105, 129). Recently, it has been proposed to focus on some clinical domain by using strategies aimed to treat both disorders simultaneously (101, 130). Besides reducing the recurrence of affective episodes, and exerting neuroprotective, mood stabilizers have been recently shown to have anti-anhedonic properties (131–134) with potential utility in the treatment of comorbid conditions (135–141). In addition, glutamatergic agents have been demonstrated to be effective in the treatment of both mood (142) and addictive disorders (82, 143); furthermore, they have been recently proposed as a valuable therapeutic option in the treatment of comorbid patients (139).

Future studies aimed at assessing brain AL in patients with BD and addiction comorbidity may help to shed light on the complex interactions underlying neurobiological vulnerability to these disorders and to improve their treatment options. Early effective treatment strategies specifically devised for comorbid patients (104, 125) could prevent, or possibly reverse, some of the neurobiological abnormalities and indicators of AL, thus potentially leading to numerous benefits for these patients.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.