Future Pharmacotherapy for Bipolar Disorders: Emerging Trends and Personalized Approaches

Abstract

Background: Bipolar disorder (BD) is a chronic and disabling psychiatric condition characterized by recurring episodes of mania, hypomania, and depression. Despite the availability of mood stabilizers, antipsychotics, and antidepressants, long-term management remains challenging due to incomplete symptom control, adverse effects, and high relapse rates. Methods: This paper is a narrative review aimed at synthesizing emerging trends and future directions in the pharmacological treatment of BD. Results: Future pharmacotherapy for BD is likely to shift toward precision medicine, leveraging advances in genetics, biomarkers, and neuroimaging to guide personalized treatment strategies. Novel drug development will also target previously underexplored mechanisms, such as inflammation, mitochondrial dysfunction, circadian rhythm disturbances, and glutamatergic dysregulation. Physiological endophenotypes, such as immune-metabolic profiles, circadian rhythms, and stress reactivity, are emerging as promising translational tools for tailoring treatment and reducing associated somatic comorbidity and mortality. Recognition of the heterogeneous longitudinal trajectories of BD, including chronic mixed states, long depressive episodes, or intermittent manic phases, has underscored the value of clinical staging models to inform both pharmacological strategies and biomarker research. Disrupted circadian rhythms and associated chronotypes further support the development of individualized chronotherapeutic interventions. Emerging chronotherapeutic approaches based on individual biological rhythms, along with innovative monitoring strategies such as saliva-based lithium sensors, are reshaping the future landscape. Anti-inflammatory agents, neurosteroids, and compounds modulating oxidative stress are emerging as promising candidates. Additionally, medications targeting specific biological pathways implicated in bipolar pathophysiology, such as N-methyl-D-aspartate (NMDA) receptor modulators, phosphodiesterase inhibitors, and neuropeptides, are under investigation. Conclusions: Advances in pharmacogenomics will enable clinicians to predict individual responses and tolerability, minimizing trial-and-error prescribing. The future landscape may also incorporate digital therapeutics, combining pharmacotherapy with remote monitoring and data-driven adjustments. Ultimately, integrating innovative drug therapies with personalized approaches has the potential to enhance efficacy, reduce adverse effects, and improve long-term outcomes for individuals with bipolar disorder, ushering in a new era of precision psychiatry.

1. Introduction

Bipolar disorder (BD) is a complex, long-lasting psychiatric illness that profoundly impacts individuals’ daily functioning and overall quality of life. It is marked by recurring episodes of mania, hypomania, and depression, affecting roughly 1–2% of people worldwide. The disorder carries a heavy burden, including high rates of disability, co-occurring medical and psychiatric conditions, and an elevated risk of suicide. Although psychopharmacological treatments, such as mood stabilizers, new generation antipsychotics, and antidepressants, are available, the long-term management of BD remains difficult. What makes this disease insidious and difficult to manage is that many patients struggle with mood swings, cognitive and functional decline, and side effects from medications. Moreover, a significant number of patients can’t achieve lasting remission, often requiring the association of multiple medications with frequent adjustments. In recent years the advances in neuroscience, molecular biology, and digital health united to a growing interest in precision psychiatry. This is an approach that combines biological, psychological, and environmental information to customize treatment for everyone, highlighting the possibility to face the limitations of current treatment approaches, by offering new possible strategies. This shift is driven by new insights into BD’s underlying biology, including immune system dysregulation, mitochondrial dysfunction, disruptions in circadian rhythms, imbalances in glutamate and GABA neurotransmission, and epigenetic changes. All this have amplified the range of potential therapies, opening doors to novel drugs and personalized interventions. Particular attention is being directed toward physiological (endo)phenotypes (stable biological traits that bridge genotype and clinical presentation), such as circadian rhythm patterns, inflammatory tone, and metabolic reactivity. These phenotypes may not only enhance patient stratification but also aid in reducing the elevated medical burden and mortality observed in BD.

This review focuses on emerging pharmacological treatments for BD, especially those currently under study or close to clinical use. Anti-inflammatory and immunomodulatory agents (like NSAIDs, cytokine blockers, and minocycline) show potential in reducing neuroinflammation linked to mood episodes. Treatments aimed at improving mitochondrial function and reducing oxidative stress, such as N-acetylcysteine, Coenzyme Q10, and PPAR agonists, address cellular energy deficits that may contribute to symptoms. Glutamatergic drugs (including ketamine, esketamine, and dextromethorphan) provide rapid relief for treatment-resistant depression, while neurosteroids and psychedelics (like zuranolone and psilocybin) offer new mechanisms that could complement traditional mood stabilizers.

Alongside drug development, digital health technologies, such as smartphone apps, wearables, and biosensors, enable continuous monitoring of physiological and behavioral signals. These tools can detect early warning signs of mood changes, improve medication adherence, and support timely interventions, paving the way for adaptive, real-time treatment adjustments. Additionally, pharmacogenomics and artificial intelligence are emerging as promising tools to predict treatment response, optimizing posology and reducing side effects.

By integrating recent advances in pharmacology, digital psychiatry, and personalized medicine, this review aims to outline possible future directions of BD management and treatment. Digital psychiatry refers to the use of digital technologies, such as smartphone-based mood monitoring, wearable sensors, digital phenotyping, and telepsychiatry platforms, to assess, predict, and manage psychiatric conditions. In BD, digital tools may enable early detection of mood episodes, real-time monitoring of medication adherence, and adaptive treatment recommendations. It is pivotal to emphasize the synergy between innovative drug mechanisms, digital monitoring, and individualized care models, as highlighted in Figure 1. While acknowledging current evidence gaps and potential challenges, the goal of this review is to show promising perspectives on translating these developments into clinical practice. Ultimately, recent research is contributing to highlight that the future of BD management lies in a comprehensive, patient-centered approach that goes beyond symptom control to promote functional recovery, resilience, and improved quality of life. In this review, personalized medicine refers to treatment strategies tailored to an individual’s biological profile. This includes pharmacogenomic approaches, such as testing for cytochrome P450 polymorphisms to guide drug metabolism and dosing (pharmacokinetics), as well as emerging methods targeting receptor-level or intracellular signaling variability via gene expression profiles (pharmacodynamics). Personalized approaches in BD may also integrate biomarkers, chronobiological patterns, and digital data streams to inform treatment selection and timing.

2. Methods

This article is a narrative review aimed at synthesizing emerging trends and future directions in the pharmacological treatment of BD, with a particular emphasis on personalized and precision-based approaches. Relevant literature was identified through a non-systematic search of electronic databases, including PubMed, Scopus, and Web of Science, covering publications from the past 15 years, with a focus on studies published in English. Keywords used in the search included: bipolar disorder, pharmacotherapy, precision psychiatry, biomarkers, neuroimaging, inflammation, circadian rhythms, glutamate, mitochondria, pharmacogenomics, personalized medicine, and novel treatments. Additional articles were selected through manual screening of reference lists from key publications and recent reviews. Inclusion criteria focused on original research articles, meta-analyses, clinical trials, and high-quality narrative or systematic reviews relevant to current or investigational treatments in BD, as well as studies addressing underlying biological mechanisms and personalized therapeutic strategies. The review was structured to highlight pathophysiological domains, emerging pharmacological targets, and technological advances supporting treatment individualization. The methodological approach prioritized the integration of clinically relevant findings and translational perspectives, rather than exhaustive coverage of all available literature.

3. Evolving Understanding of Bipolar Pathophysiology

3.1. Genetic and Epigenetic Insights

BD has a high hereditary component as classically demonstrated by studies of relatives and twins (approximately 80% of the risk is due to hereditary factors [1]). In addition, although it is not possible to predict whether the disorder will be inherited, children of bipolar patients are 7 times more likely to develop the disorder [2]. Studying individual chromosomes, genes and loci in relation to bipolar disorder is complicated by the disorder’s genetic and phenotypic variability, and by its genetic and pathophysiological overlap with other conditions such as major depressive disorder, schizophrenia and schizoaffective disorder. Despite the additional difficulty also given by genetic polymorphisms, genomewide association analysis (GWAS) studies using very large samples have correlated some genes with BD, contributing to heritability by highlighting associations between the disease and genetic variants, e.g., DGKH, CACNA1C, ANK3 [3]. For instance, in a 2016 study, the sample consisted of nearly 10,000 bipolar disorder patients and 30,000 controls, finding an association between BD and two loci added to the growing list of variants: an intergenic region on 9p21.3 and markers within ERBB2 [4] In another 2019 genomewide association analysis study in which more than 50 thousand people including cases and controls were studied, 30 loci associated with BD were identified, including 20 novel ones. In 2021, a study was conducted in which 64 associated genomic loci were highlighted by sampling 41,917 bipolar disorder cases and 371,549 controls of European ancestry. In particular, this study found 33 new genome-wide significant loci, including one related to major histocompatibility complex (MHC) [5]. Genes contained in the loci encode ion channels, neurotransmitter transporters and synaptic components [6] Regarding chromosomes, significant associations were found with 2p, 4p, 4q, 6q, 8q, 11p, 12q, 13q, 16p, 16q, 18p, 18q, 21q, 22q and Xq [7], while a further study had added the 22q area, on which there is the catechol-o-methyltransferase gene that could be connected to BD, as well as the serotonin transporter gene on 17q [8]. Other genes studied for a possible role in BD are COMT, DAT, HTR4, DRD4, DRD2, HTR2A, 5-HTT, the G72/G30 complex, DISC1, P2RX7, MAOA and BDNF [9]. Many of these genes, such as COMT, DRD2, HTR2A, MAOA, and 5-HTT, are included in commercially available pharmacogenomic panels that clinicians can order for psychiatric practice, with results typically available within days. For clinicians interested in incorporating pharmacogenomic testing into their practice, several CLIA-certified and FDA-cleared companies offer ready-to-use kits (typically saliva- or buccal-swab based) that can be easily collected in outpatient settings. Results are provided via secure online platforms, often with embedded clinical decision support for drug selection based on pharmacokinetics and pharmacodynamics. Many providers also assist with insurance processing or offer transparent out-of-pocket pricing for patients.

However, other genes (e.g., P2RX7, G72/G30) are not currently part of routine clinical testing and remain investigational. Access to testing may vary by country, but in many cases, psychiatrists can initiate testing directly through certified providers without requiring external laboratory coordination by the patient.

Advances in genomic technologies are increasingly informing the development of personalized treatment approaches in psychiatry. While most genetic studies in BD have focused on disease susceptibility, pharmacogenomic testing, particularly of genes involved in drug metabolism (e.g., CYP2D6, CYP2C19) and receptor function (e.g., DRD2, HTR2A, BDNF), holds promise for optimizing pharmacotherapy. Commercial testing kits now enable clinicians to assess individual metabolizer status and predict responses or side effect risks, though widespread clinical adoption is still evolving.

The presence of pharmacogenomic test results in a patient’s medical record may carry legal and ethical implications. If such results indicate a clinically significant metabolism issue (e.g., CYP2D6 poor metabolizer) and are not considered in treatment selection, especially in the event of an adverse outcome, institutions may be exposed to potential liability. As pharmacogenomic integration increases, hospitals and clinics are encouraged to develop clear policies and education protocols to ensure that test results are meaningfully reviewed and documented in the prescribing workflow [10].

It is pivotal to outline that the use of pharmacogenomic data in prescribing is not yet mandated in most jurisdictions, and legal standards are evolving. However, in cases where pharmacogenomic results are available and indicate that a patient is a poor metabolizer (e.g., CYP2D6) of a specific drug, physicians are encouraged to document their clinical rationale for prescribing decisions that deviate from pharmacogenomic-based recommendations. Although liability would depend on country-specific malpractice frameworks and clinical circumstances, failure to consider actionable pharmacogenomic data, especially if harm results, could increasingly be viewed as a deviation from emerging standards of care. Future studies integrating drug-gene pairs with clinical phenotypes and biomarkers may enhance treatment selection and reduce adverse outcomes. Genomic-guided prescribing represents an important emerging tool in the broader framework of precision psychiatry.

3.2. Neuroinflammation and Immune Dysregulation

Chronic neuroinflammation and immune dysregulation is another mechanism that affects the onset of bipolar disorder [11,12]. For example, glucocorticoids, which normally during acuity in times of stress are released with the beneficial effect of increasing metabolism and cognitive function, are decreased in chronic stress due to the continuous stimulus [13]. Changes also occur at the level of receptor binding: glucocorticoids by binding to receptors with high affinity promote structural changes such as dendritic growth in the prefrontal cortex, hippocampus and amygdala that are beneficial to working and long-term memory [14]. Glucocorticoid receptors are reduced under chronic stress conditions, which is associated with a reduction in cortisol sensitivity and negative consequences for neuronal function [15,16]. There is a compensatory mechanism called allostasis that allows the body to adapt to stress with physiological responses [17]. When stress is chronic, repetitive, and frequent, an allostatic load is created, which could lead to disease progression through organic damage [18]. This hypothesis of allostatic load transposed to the brain gave rise to the concept of neuroprogression, which adapted to bipolar disorder gives an explanation for the fact that through repeated episodes of mania and depression the brain undergoes progressive changes and damage [19,20]. Repeated episodes then accelerate the progression of the disease through increased vulnerability to stress and reduced resilience [21]. This gives rise to impairment of cognitive functions such as memory and decision-making and emotional processes such as mood regulation [22,23]. Biological mechanisms involved include neuroinflammation and mitochondrial dysfunction through activation of proinflammatory responses by the production of cytokines whose receptors are present on brain cells and mitochondria (IL-6 IL-10 TNF a), cellular and mitochondrial damage by increased calcium in cells, and oxidative stress [24]. Intracellular Calcium entry, promoted by increased expression of L-type calcium channels in response to stress, opens pores in mitochondria resulting in cytochrome c release and cell death (apoptosis) via caspase-3, flanked by apoptosis via caspase-8 (extrinsic pathway of apoptosis via receptors such as TNF) [25]. This causes a decrease in Adenosine Triphosphate (ATP) production [26]. Oxidative stress is a consequence of mitochondrial apoptosis, which produce free radicals that damage neurobiological structure (lipids, proteins, neuronal DNA) [27]. All this exacerbates neurobiological damage, with neuronal death due to decreased capacity for neuronal repair and mitochondrial transport and to increased frequency of manic and depressive episodes and their severity [28].

Pro-inflammatory cytokines (IL-1β, IL-6, TNF-α) play a role in the activation of immune cells (such as monocytes, macrophages, and microglia) and at the same time their increase is related to the activation of microglia [29]. Under physiological conditions, microglia is involved in synaptic pruning, remove dead cells and debris, and release brain-derived neurotrophic factor (BDNF), that promotes neuronal growth, in addition to the fact that at the time when there is trauma or pathogen damage they regulate neuroinflammation [30]. They are derived from the myeloid lineage but there are two different phenotypes into which microglia and brain macrophages can differentiate: classically activated (M1) and alternatively activated (M2) populations [31]. In the M1 group are cells with pro-inflammatory function that release cytokines such as TNF-α and IL-6 and IL-1, which together with oxidative metabolites, with the function of eliminating the trigger that activated them but thereby causing increased cellular damage. The second group includes cells with an anti-inflammatory function that modulate the former, promoting tissue repair and healing [32]. The switch from M1 to M2 is therefore very important in resolving inflammation, which may not be the case in states of chronic inflammation where there may be a prevalence of M1 cells resulting in neuronal dysfunction or neuronal degeneration [33]. In patients with BD, an excess of M1 or a decrease in M2 has been observed in the manic phase, with the possibility of correlating symptomatology with M1/M2 imbalance [34]. It can be observed that in vitro responses to activated microglia could be very different in vivo due to the absence in in vitro models of inhibitory factors [35,36]. This neuroinflammatory response, leading to injury, may affect the progression of BD [37].

Inflammation, however, appears to be multisystemic, having studies found via biomarkers of inflammation and oxidative stress in patients with acute onset of BD higher systemic toxicity than healthy controls [38,39]. Further studies are necessary to identify peripheral biomarkers and put them in a causal relationship with systemic toxicity [40]. There could be specific biomarkers for different phases of the disease (active or remission periods), and some have been linked to BD allostasis and neuroinflammation [41]. There are currently no biomarkers for early diagnosis of BD but there are three promising categories among which we find imaging signs, genetic loci, and metabolic molecules [42,43].

3.3. Mitochondrial Dysfunction and Oxidative Stress

Neurons are cells that require a high level of energy, and this energy is produced by mitochondria in the form of ATP [42]. The energy is used for neurite growth, calcium homeostasis and signaling, the maintenance of ion gradients and electrical signaling, and the long-distance axonal transport [44]. The gray matter requires three times the energy of the white matter, and altogether the brain consumes 20–25% of the human body’s oxygen and glucose, respectively [45]. There are some areas of neurons that require more energy than others: presynaptic and postsynaptic endings, active growth cones or axonal branches, and nodes of Ranvier [46]. In BD and other psychiatric disorders, the number of mitochondria is constantly changing, producing their altered distribution and problems in transport. Stress and axonal trauma result in altered movement of mitochondria in axons, which is essential for synaptic remodeling [47]. Mitochondrial dysfunction and reduced metabolism can lead to altered synaptic plasticity, which is correlated with psychiatric issues such as BD and schizophrenia [48]. In addition to the altered distribution of mitochondria and alterations in the level of metabolites, differences in mitochondrial morphology between controls and BD patients also emerged in the postmortem prefrontal cortex, primary fibroblasts and lymphocytes [49]. It has been shown that people with mitochondrial disease are 20 times more likely to develop symptomatology compatible with BD. BD-like behavior has also been shown in mouse models in which mutations of mitochondrial genes and mutations and deletions of mitochondrial DNA have been found in some patients with BD [50].

Other alterations found in BD include changes in the levels of enzymes involved in the Krebs cycle. This cycle is essential for producing molecules (such as NADH and FADH₂) that fuel oxidative phosphorylation (OXPHOS) in mitochondria. This can cause energy deficit in neurons and increased oxidative stress [51]. In addition, increased free radical production and reduced antioxidant capacity are observed in BD patients, contributing to mitochondrial dysfunction and neuroinflammation, induced by ROS that stimulate the production of inflammatory cytokines such as IL-1β, TNF-α [52]. It is relevant to highlight that the brain is very vulnerable due to its high content of lipids that are easily oxidized, high oxygen consumption, and the presence of few natural antioxidants [53]. Mitochondrial dysfunction could increase the burden of oxidative stress, which in BD could worsen the disease, so improving mitochondrial functions could be a way to pursue to treat BD [54,55] Specifically, oxidative stress causes problems in signal transduction, on neuronal plasticity and cellular resilience. The activity of antioxidant enzymes CAT, SOD and GSH-Px, NO and GSH levels, and lipid peroxidation are found to be altered in BD, as well as DNA damage [56]. However, discordant results were found on the association between antioxidant enzymes or oxidative stress parameters and psychiatric pathology or its clinical phase [49]. In a meta-analysis [57] of 27 studies (971 patients) showed that lipid peroxidation, DNA/RNA damage, and nitric oxide (NO) are increased in all phases of BD, in a 2010 study mitochondrial proteins in BD patients are more oxidized, while another meta-analysis shows that lipid peroxidation levels might be higher during manic or depressive episodes than during remission and in a 2016 study NO levels show increased even in remission [58,59,60].

Research demonstrating changes in serum antioxidant enzymes at different phases of BD has yielded mixed results, suggesting their limited reliability as biomarkers. Importantly, the current literature often conflates distinct categories of biomarkers: (1) trait biomarkers, indicating a predisposition or chronic underlying feature of BD; (2) predictive biomarkers, which may signal risk before illness onset; and (3) state biomarkers, which fluctuate according to mood episodes. Many immune and oxidative stress markers, such as IL-6, are highly sensitive to external influences like circadian rhythm disruptions, metabolic comorbidities, and current affective state. This biological variability complicates their interpretation and limits clinical utility. To improve translational impact, future research should aim to clearly define the biomarker type under investigation and control for dynamic physiological variables [61]. An important factor conditioning neuroplasticity in BD is BDNF. Indeed, it appears that BDNF levels are lower in the depressive and manic phases than in the euthymic phase and decrease as the disease progresses. BDNF also decreases in relation to stress and trauma [62]. It has a role in regulating the release of glutamate, serotonin, and GABA [63]. In addition, it is hypothesized that in patients with BD the levels of BDNF and molecules that regulate oxidative stress are altered, although there are many contradictory results probably due to various factors such as metabolic comorbidities (obesity, diabetes), differences in clinical status (manic phase, depressive phase, euthymia/remission) and pharmacological treatments [64,65]. Alterations in BDNF levels and antioxidant markers vary widely but confirm that BD is associated with an imbalance between oxidative damage and antioxidant defenses [66].

3.4. Neurotransmitter and Receptor System Imbalances

The glutamatergic system plays a central role in the pathophysiology of BD; other important factors include the gabaergic system, monoaminergic system, glia-neuron interactions and endocrine dysfunction [67]. Monoamines have been related to both Major Depressive Disorder and BD. It was found that elevated CSF levels of 3-methoxy-4-hydroxyphenylglycol (MHPG), a metabolite of norepinephrine, were related to agitation and anxiety in depressed patients, in a study that included depressed and bipolar patients [68]. In addition, there was a decrease in MHPG in the CSF of suicidal bipolar subjects compared with controls and immunoreactivity of locus coeruleus processes [69,70]. The change in MHPG levels could be useful to indicate phase switch in BDI and to direct the best treatment [71]. Abnormal expressions of several NE- and 5-HT-related genes were shown in BD patients compared with controls, data from the Stanley Medical Research Institute online genomics database (SMRI) [72]. As for dopamine, there is still no conclusive evidence linking it to BD. However, pharmacological and imaging studies have been done to hypothesize its role in the disease.

The dopaminergic hypothesis suggests that the BD phases may depend on an altered balance between dopamine transporters and receptors. During mania phases, there could be an increased availability of D2/D3 receptors and a hyper-responsive reward system; in the depressive phase, on the other hand, there would be an increase in the dopamine transporter, causing its over-reuptake and reducing the amount available [73]. Excessive dopaminergic activity in a BD patient could lead to down-regulation of dopamine receptors, facilitating the transition to a depressive state [74]. A study supporting the dopaminergic dysfunction theory correlates tardive dyskinesia, even in patients with BD who do not have drug therapy, to severe bipolar symptoms [75]. Regarding serotonin (5HTT), imaging data showing a correlation between its transporter and BD are conflicting, reporting sometimes an increase in monoamine binding and sometimes a decrease in various brain areas [76,77]. Serotonergic System dysfunction seems to be linked to BD and its severity, nevertheless the 5HT1A receptor binding has not been considered a useful predictor of treatment response [78]. Regarding GABA, it was shown that platelet uptake of GABA in BD patients in the depressive phase was increased, while it was decreased in the manic phase [79]. Glutamate, on the other hand, had increased platelet uptake during mania. The depressive and manic phases were more severe in correlation with altered uptake of GABA and glutamate. Through the study of the enzyme involved in GABA synthesis, GAD, a difference in GABAergic transmission was noted at the level of various brain areas in patients with unipolar and bipolar depression [80]. Regarding glutamate, studies have shown in bipolar subjects a decreased expression of the NR1 and NR2A subunits of N-methyl-D-aspartate (NMDA) receptors in the hippocampus and an increase in the glutamate transporter VGluT1 in the anterior cingulate cortex [81]. Neurons release glutamate, which is taken up by glial cells that return it to neurons in the form of glutamine, which is useful for synthesizing other neurotransmitters. In several neurological diseases, high levels of glutamate and glutamine are associated with cognitive problems [82]. Gene expression studies provide further support for neurotransmitter system dysregulation in BD. Postmortem analyses of the prefrontal cortex and anterior cingulate cortex have reported reduced expression of dopamine D2 receptor (DRD2) and D3 receptor (DRD3) genes in BD patients compared to controls [83,84]. In addition, GABAergic dysfunction has been linked to decreased expression of GAD1 (glutamic acid decarboxylase 67), an enzyme essential for GABA synthesis, and GABRA1/GABRB2 (GABA-A receptor subunits) in BD brains. These findings support receptor-level imbalances as part of BD pathophysiology, even in medication-free or minimally treated samples.

Studies on rats have demonstrated that mood stabilisers can cause changes in mood by modulating glutamate receptors. In addition, postmortem studies have shown that excessive neuronal activation due to glutamate can result in excitotoxicity at the frontal cortex level in BDs, abnormal excitatory synaptic connections, and altered glutamate receptor function at pre- and post-synaptic levels. In vivo TMS (transcranial magnetic stimulation) studies showed reduced cortical inhibition, probably related to glutamatergic hyperactivity. Extra synaptic ionotropic receptors are activated by excess glutamate, causing calcium influx and Ros production resulting in neurotoxicity [85]. Neuroimaging studies have revealed quantifiable alterations in neurotransmitter systems in individuals with bipolar disorder. For example, elevated dopamine D2/D3 receptor binding potential has been reported in manic patients compared to healthy controls, using PET with [¹¹C]-raclopride, suggesting heightened dopaminergic transmission [86]. In parallel, proton magnetic resonance spectroscopy (¹H-MRS) studies have shown increased glutamate and Glx concentrations in the anterior cingulate cortex of symptomatic bipolar patients compared to controls, supporting glutamatergic dysfunction [87,88]. These findings reinforce the hypothesis of neurotransmitter system imbalance as a core pathophysiological feature of BD.

4. Future Pharmacotherapy: Emerging Drug Classes and Mechanisms

4.1. Anti-Inflammatory and Immunomodulatory Agents

Immune system dysregulation has been increasingly recognized as a key factor in the pathophysiology of BD. Based on the findings of a recent systematic review [89], pharmacological treatments for BD can be divided into two categories according to their effects on inflammatory biomarkers. The first group includes agents such as mood stabilizers (e.g., lithium), antipsychotics (e.g., quetiapine), antidepressants (e.g., ketamine), and antibiotics (e.g., minocycline), which appear to modulate immune responses by increasing both pro-inflammatory (e.g., IL-6, TNF-α) and anti-inflammatory (e.g., IL-4, IL-8, IL-10) cytokines. The second group includes medications like memantine, dextromethorphan, infliximab, NSAIDs (aspirin, celecoxib), antidiabetics (pioglitazone), and omega-3 fatty acid supplementation, which tend to reduce inflammatory markers (such as TNF-α, IL-6, IL-1β, and CRP) and/or increase levels of brain-derived neurotrophic factor (BDNF) in patients with BD. Within the first group, lithium has shown neuroprotective effects in both neurodegenerative and psychiatric conditions, with potential anti-inflammatory, antioxidant, and anti-apoptotic properties [90,91]. Its mechanism of action is thought to involve inhibition of glycogen synthase kinase 3 (GSK3), though its precise role in mood stabilization remains unclear [92,93]. However, studies suggest some controversy, as they indicate increases in pro-inflammatory factors (IL-6, MCP-1, IFN-γ, TNF-α), but, at the same time, they report increases in anti-inflammatory cytokines such as IL-4, IL-8, and IL-10. Similarly, quetiapine has demonstrated anti-inflammatory properties in preclinical studies, but clinical data remains mixed, since associations between quetiapine use and increased levels of pro-inflammatory cytokines such as IL-6 and TNF-α have been described [81]. Of particular interest is evidence showing that the combination of lithium and quetiapine resulted in lower levels of inflammatory markers (TNF-α, TGF-β1, IL-17, IL-23), suggesting that certain drug combinations may exert synergistic immunomodulatory effects [94]. Despite growing evidence for immune involvement in BD, these findings cannot be fully interpreted without considering individual differences in chronobiological regulation. Dysregulation in cortisol and melatonin secretion, key circadian markers, may significantly modulate immune activity and inflammatory tone. Future studies should examine immune response profiles in the context of individual sleep-wake patterns, hormonal rhythms, and chronotypes. This approach would allow integration of immunological data with cortisol/melatonin dynamics and refine treatment strategies. Importantly, pharmacological research in BD should move away from reliance on average values from large, heterogeneous samples. Instead, efforts should focus on precise phenotypic stratification, based on biological, behavioral, and chronobiological parameters, which may better predict treatment response and disease progression.

It is also important to reconsider traditional outcome measures in clinical trials involving immunomodulatory agents. The primary aim in such studies may not always be immediate symptom improvement, but rather biological modulation of identified dysregulations, such as reducing systemic inflammation or correcting aberrant immune-circadian rhythms. These intermediate outcomes, if mechanistically linked to illness trajectory, may serve as meaningful short-term goals. As such, inclusion criteria for these trials should be phenotype-specific, targeting patients with elevated inflammatory markers, for example, rather than relying solely on syndromic diagnoses or symptom severity scores. This approach could enhance the translational value of precision-targeted interventions

Ketamine, an NMDA receptor antagonist, is believed to exert its effects via mechanisms involving the kynurenine pathway and mTOR signaling [95]. Although both animal and human studies indicate that ketamine may reduce pro-inflammatory cytokines [96,97,98], data suggest it may increase levels of both IL-6 and IL-8 [99]. Notably, studies reported no significant changes in cytokine levels post-treatment and no correlation with mood symptom improvement, adding to the complexity of its immunomodulatory profile.

Minocycline, an antibiotic with putative anti-inflammatory properties, has yielded mixed results. While an open-label study showed symptom improvement in BD patients and reported reduced levels of IL-12p70 and CCL26, it also revealed increased IL-12/23p40 levels, with no significant changes in most other cytokines assessed [100]. Additionally, another trial combining minocycline with aspirin indicated a stronger clinical response in patients with elevated IL-6 levels at baseline, and a significant reduction in IL-6 was associated with symptom improvement, while C-reactive protein (CRP) levels did not show meaningful interaction with treatment [101].

Memantine and dextromethorphan, two antidepressants with mechanisms similar to ketamine, appear to exert neuroprotective and neurogenic effects via BDNF/TrkB signaling [102,103]. Their anti-inflammatory actions, supported by prior studies, align with findings from this review. Memantine, in particular, has shown consistent reductions in IL-6, IL-1β, TNF-α, IL-8, and CRP levels across multiple studies [104,105,106,107]. These changes were often accompanied by clinical improvements in depressive and manic symptoms and correlated with reduced Hamilton Depression Rating Scale (HDRS) and Young Mania Rating Scale (YMRS) scores. Increases in BDNF levels were also observed in memantine-treated patients [104,105], further supporting its neurotrophic potential.

Dextromethorphan, when added to valproic acid, was associated with significant improvements in manic symptoms and increased BDNF levels, although no direct correlation between BDNF and clinical improvement was found [105,108]. Reductions in CRP were also reported when dextromethorphan was combined with memantine [105], whereas no significant changes were observed in TNF-α and IL-8 levels [108].

Other agents also showed anti-inflammatory potential. Infliximab, a monoclonal antibody used in autoimmune diseases, reduced CRP, TNF-α, and sTNFR2 levels, consistent with its TNF-α–targeted mechanism [109,110,111], although no antidepressant effect was found. Notably, cognitive and anhedonic symptoms improved in some studies, suggesting a potential for selective symptom modulation.

Low-dose aspirin has demonstrated modest adjunctive efficacy in reducing depressive symptoms in bipolar disorder. For example, Nery et al. [112] reported that aspirin, when combined with N-acetylcysteine, produced significant symptom improvement compared to placebo in bipolar depression. Similarly, Savitz et al. [113] noted reductions in inflammatory markers and mood symptoms in BD patients treated with aspirin and minocycline. However, these findings are limited by small sample sizes and short durations. In contrast, ibuprofen lacks in consistent antidepressant effects in BD and is not recommended for clinical use in this context. Further large-scale RCTs are needed to establish whether nonsteroidal anti-inflammatory drugs NSAIDs can play a meaningful therapeutic role in bipolar depression.

Celecoxib, particularly when combined with escitalopram, was associated with lower CRP levels and non-significant reductions in IL-1β, with some evidence of improved depressive symptoms in treatment-resistant BD patients [114,115,116,117]. Some trials failed to demonstrate differences in inflammatory markers compared to placebo [114], indicating heterogeneity in the response. Moreover, when added to Electroconvulsive Therapy (ECT), celecoxib reduced TNF-α but not IL-1β, IL-6, or CRP [118].

Similarly, pioglitazone, a PPARγ agonist used in diabetes, showed reduced IL-6 and CRP levels in BD patients after eight weeks of treatment [119], and the reduction in IL-6 was associated with improvement in depressive symptoms.

Additional studies have proposed alternative anti-inflammatory interventions. Aldesleukin (low-dose IL-2) showed antidepressant effects and increased sIL-2RA and CRP levels without altering BDNF or IL-7 [120]. Omega-3 fatty acids, administered over two months, were linked to reductions in IL-6, TNF-α, and CRP, which correlated with improvements in depressive symptoms [121]. In contrast, probiotic supplementation showed no significant changes in inflammatory markers or clinical outcomes [122].

Although a growing body of evidence supports immune dysregulation in BD, clinical trials targeting the immune system have so far yielded mixed or disappointing results, both in psychiatry and neurology. As seen in Alzheimer’s disease, the most promising outcomes have stemmed from multimodal lifestyle interventions, rather than pharmacological immune suppression [111]. In BD, it remains unclear whether inflammatory markers such as IL-6 are causally linked to core pathology or are secondary to factors such as disrupted circadian rhythms, chronic stress, or metabolic dysregulation. Moreover, immune alterations often co-occur with mood changes, but their correction does not consistently yield symptom relief. These considerations argue for a cautious, hypothesis-driven approach to immunomodulatory treatments, ideally within a broader framework incorporating chronobiological and behavioral targets. Among emerging targets, the P2X7 receptor (P2RX7), a purinergic ligand-gated ion channel expressed on microglia, has drawn increasing attention. P2X7 activation is associated with proinflammatory cytokine release and neuroinflammation, both implicated in BD pathophysiology. Genetic variants in P2RX7 have been linked to prospective clinical outcomes in BD, and selective P2X7 antagonists are under investigation in preclinical models and early-phase trials [123]. Although clinical evidence remains limited, this receptor represents a promising avenue for targeted immune modulation in BD.

4.2. Mitochondrial Enhancers and Oxidative Stress Modulators

Recent advances in the understanding of BD have brought growing attention to the role of mitochondrial dysfunction and oxidative stress in its pathophysiology. These insights open new avenues for personalized treatment strategies that address psychiatric symptoms alongside metabolic imbalances, with the ultimate goal of developing precision medicine tailored to individual neurobiological profiles [124]. A key concept emerging from current literature is the need to modulate mitochondrial activity depending on the mood state, enhancing mitochondrial function during depressive episodes and potentially attenuating it in manic states, or to bolster antioxidant defenses to prevent damage caused by reactive oxygen species (ROS), whether due to increased cellular respiration or insufficient intrinsic antioxidant activity [125]. Importantly, mitochondrial-targeted agents often exhibit a delayed therapeutic onset compared to conventional antipsychotics or mood stabilizers and are thus typically explored as adjunctive treatments rather than monotherapies [126].

In this context, Kuperberg et al. [124] provides a comprehensive overview of several mitochondrial enhancers and oxidative stress modulators under investigation for BD. One notable class is that of PPAR agonists, targeting the peroxisome proliferator-activated receptors (PPARα, PPARδ, and PPARγ), which are nuclear receptors involved in regulating energy homeostasis, oxidative stress response, mitochondrial biogenesis, and inflammatory processes. In the central nervous system, PPARs also influence excitatory neurotransmission and myelination [127,128]. While originally developed for metabolic disorders like hypertriglyceridemia and type 2 diabetes, PPAR agonists have shown neuropsychiatric potential. Of particular interest is PGC-1α, a coactivator of PPARs weakly associated with lithium response [129]. Among the selective PPARγ agonists, pioglitazone has demonstrated antidepressant efficacy, though its clinical use is constrained by adverse effects such as weight gain, edema, and increased cardiovascular risk [130]. Bezafibrate, a pan-PPAR agonist targeting all three isoforms, appears more promising due to its favorable safety profile and is currently under clinical evaluation for BD. It promotes mitochondrial biogenesis and exhibits antioxidant and anti-inflammatory effects [131]. The mechanisms of action of minocycline, a tetracycline antibiotic with neuroprotective properties relevant to mitochondrial dysfunction in BD, include scavenging ROS, restoring glutathione (GSH) levels, modulating glutamatergic neurotransmission, and regulating apoptosis through upregulation of the anti-apoptotic gene Bcl-2, directly impacting mitochondrial apoptotic pathways [132,133]. Although promising effects, such as symptom improvement and increased gray matter volume, have been observed in schizophrenia, results in mood disorders have been inconsistent [134].

N-acetylcysteine (NAC) exerts multifaceted actions including antioxidant, anti-inflammatory, and glutamate-modulating effects. By supplying cysteine for GSH synthesis, NAC enhances mitochondrial resilience against oxidative stress while also restoring respiratory complex activity in preclinical models [135,136]. Clinical studies suggest a delayed antidepressant effect, often manifesting only after prolonged treatment. In one trial, improvement in depressive symptoms emerged 20 weeks after drug discontinuation in a group receiving NAC in combination with other mitochondrial agents, suggesting either a protracted therapeutic effect or post-treatment rebound [137]. Interestingly, a transient increase in manic symptoms was reported at week 4 in some patients, raising concerns about potential pro-manic effects of enhanced mitochondrial biogenesis [137].

Coenzyme Q10 (Co-Q10), or ubiquinone, is a crucial cofactor in the electron transport chain (ETC), essential for ATP production. It also has antioxidant and anti-inflammatory properties, supports membrane stabilization, regulates apoptosis, and mitigates glutamate-induced excitotoxicity [138,139]. Endogenous Co-Q10 synthesis declines with age, and its deficiency has been linked to psychiatric and neurological conditions [140,141]. In BD, adjuvant Co-Q10 supplementation has been associated with reductions in depressive symptoms, although the therapeutic effects typically appear after at least 8 weeks of treatment [142,143].

Finally, melatonin, a hormone primarily known for circadian regulation, is emerging as a mitochondrial modulator. It enhances mitochondrial function by increasing expression and activity of electron-transport chain proteins, promoting mitochondrial biogenesis, stabilizing mitochondrial membranes, and protecting mitochondrial DNA [144,145]. Melatonin also upregulates antioxidant gene expression and GSH production. Patients with BD have been found to have decreased melatonin levels, with state-dependent variation, and circadian rhythm disturbances are common in this population [146,147]. Clinical trials with melatonin receptor agonists such as ramelteon suggest potential utility in preventing depressive relapses, although results across studies remain mixed [147].

Collectively, these findings underscore the potential of mitochondrial enhancers and oxidative stress modulators as adjunctive strategies in BD treatment, though their delayed onset of action and complex pharmacodynamics require careful clinical consideration.

4.3. Glutamatergic Modulators

Glutamatergic system dysfunction has been implicated in the pathophysiology of mood disorders, including bipolar depression, and has become a central target in the development of rapid-acting antidepressants (RAADs). Among these, the NMDA receptor antagonist (R,S)-ketamine has shown the most consistent and compelling evidence of efficacy. Several placebo-controlled, double-blind, randomized trials have demonstrated that a single subanesthetic-dose infusion of (R,S)-ketamine exerts rapid and robust antidepressant effects in individuals with bipolar depression [148,149]. A meta-analysis of nine randomized, placebo-controlled studies found that these effects typically emerge within 40 min post-infusion, peak at around 24 h, and taper off by 10–12 days [43]. Repeated infusions have been shown to prolong these antidepressant effects [150,151].

Additional compounds targeting the glutamatergic system have also been evaluated. Dextromethorphan, a non-selective, non-competitive NMDA receptor antagonist with activity at opioid and sigma-1 receptors, was tested adjunctively with valproic acid in a 12-week randomized, placebo-controlled trial involving 250 individuals with bipolar depression but showed no significant antidepressant effects [152]. Conversely, Nuedexta, a combination of dextromethorphan and quinidine, which inhibits CYP2D6 and enhances Central Nervous System bioavailability, demonstrated clinical benefit in a retrospective chart review of 77 individuals with BD-II or BD-NOS, showing significant improvements in Clinical Global Impression (CGI) scores over a 90-day period [153].

4.4. Neurosteroids

Neuroactive steroids, such as allopregnanolone and its synthetic analogs, represent a novel and promising class of therapeutic agents due to their modulatory action on GABA-A receptors, particularly at δ-subunit-containing extrasynaptic sites [154,155]. These compounds, derived from progesterone or synthesized de novo in the brain, influence neurodevelopment, neuroplasticity, and neurotransmission, and have shown protective effects in acute stress while being reduced in chronic stress and depression [156,157]. Zuranolone (SAGE-217), a next-generation oral analogue of allopregnanolone, was developed as a rapid-acting, short-course treatment for mood disorders. Compared to its intravenous predecessor brexanolone, approved for postpartum depression, zuranolone offers greater practicality in both clinical and outpatient settings due to its oral formulation. In BD, zuranolone is under investigation for its antidepressant effects. While previous Maior Depressive Disorder trials showed transient benefits [158], sustained improvement in bipolar depression suggests zuranolone may be a promising adjunctive or alternative treatment, pending confirmation in controlled studies [159,160,161].

4.5. Psychedelics and Novel Rapid-Acting Antidepressants

Over the past two decades, psychedelics have re-emerged as promising treatments for a range of psychiatric disorders. Among these, psilocybin has garnered particular interest for its rapid-acting antidepressant effects. Psilocybin is a naturally occurring compound found in mushrooms of the Psilocybe genus, whose active metabolite, psilocin, acts primarily as a potent 5-HT2A receptor agonist [162,163]. This receptor activity is thought to contribute to the drug’s ability to disrupt maladaptive brain network connectivity, particularly in circuits implicated in rigid negative thought patterns and self-referential processing [164]. Psilocybin-assisted therapy typically involves one or two high-dose sessions embedded within a framework of preparatory and integrative psychotherapy. In one trial, psilocybin showed comparable effects to escitalopram, although the absence of a placebo arm limited efficacy conclusions [165]. Another study reported significant reductions in depressive symptoms compared to a waitlist control [166]. However, it remains unclear whether the observed therapeutic effects are primarily pharmacological or result from the synergy between drug action and psychotherapeutic support. A recent study by Rosenblat et al. (2024) [167] represents a notable advancement in this field, exploring the safety and feasibility of psilocybin-assisted psychotherapy (PAP) in individuals with bipolar II disorder (BDII), a population historically excluded from psychedelic trials. The study enrolled patients with ultra-refractory depression, including BDII, characterized by high rates of comorbidity and suicidality. Notably, no treatment-emergent mania, hypomania, or psychosis was observed in BDII participants, suggesting that PAP may be safely administered under controlled conditions in complex clinical populations. The study also utilized a flexible dosing strategy, allowing for additional psilocybin sessions in cases of symptom recurrence, which proved both feasible and safe. Improvements in depressive symptoms were consistent with prior findings, although the anxiolytic effects appeared less robust.

Another line of research into rapid-acting antidepressants has focused on cannabinoids, particularly cannabidiol (CBD). CBD has shown anxiolytic and antidepressant-like effects in preclinical models, likely mediated through 5-HT1A receptor facilitation and CB1 receptor modulation [168,169]. A pilot randomized controlled trial assessing adjunctive CBD in acute bipolar depression (BDI and BDII) found no significant difference compared to placebo in the overall group. However, a sensitivity analysis indicated that higher doses (300 mg/day) might benefit non-early responders [170]. Importantly, CBD was well tolerated and not associated with manic switching, highlighting its potential as a mood-stabilizing adjunct.

In recent years, ketamine and its S-enantiomer, esketamine, which had previously been highlighted for their glutamatergic mechanisms, have emerged as promising treatments due to their rapid antidepressant and anti-suicidal effects, particularly in individuals resistant to conventional therapies [171]. These compounds are among the first antidepressants capable of alleviating depressive symptoms and suicidal ideation within hours in many patients [172]. Intravenous ketamine is a racemic mixture containing both R- and S-enantiomers, which share overlapping mechanisms of action, mainly through antagonism of the NMDA receptor and also activity at the σ1 receptor [173]. More recently, intranasal esketamine has been approved for treatment-resistant depression (TRD), offering a more convenient alternative to intravenous infusions, especially for patients with bipolar disorder who have not responded to traditional treatments [174,175]. The intranasal route facilitates clinical use, requiring no infusion and allowing for weekly or biweekly dosing after an initial twice-weekly induction phase [176].

Esketamine is currently FDA approved for adults with unipolar TRD, used in combination with oral antidepressants, and for major depressive disorder with suicidal ideation or behavior [177,178]. Although not officially approved solely for suicidal ideation, multiple studies support the use of intranasal esketamine in patients with major depressive disorders experiencing acute suicidal thoughts or actions.

The systematic review by Nunez and colleagues [179] examined eleven studies involving patients with bipolar depression, focusing on the efficacy and tolerability of intravenous (IV) racemic ketamine, both as single and repeated infusions. The review found that both approaches were similarly effective and generally well tolerated. However, several studies noted a return of depressive symptoms shortly after the final infusion [180,181,182], raising concerns about the limited duration of ketamine’s therapeutic effects and the need for long-term treatment strategies. This underlines the importance of establishing clear dosing protocols for maintenance treatment and improving patient phenotyping to identify those most likely to benefit.

Studies on intranasal and subcutaneous esketamine demonstrated similar efficacy in both TRD and treatment-resistant bipolar depression (TRBD), with additional benefits on anhedonia [175,183]. Although serial infusions tended to show slightly better outcomes than single infusions, these differences were not statistically significant and may reflect variations in study design, such as more stringent inclusion criteria or lower expectancy effects in randomized controlled trials.

There was considerable heterogeneity across the studies, especially in patient characteristics and ketamine dosing, though most utilized a subanesthetic dose. Some evidence suggested a greater antidepressant response in patients with BDII compared to BDI [184]. Additionally, several studies highlighted ketamine’s rapid anti-anhedonic effects and its potential to reduce suicidal ideation independently of overall depressive symptom improvement [185,186,187,188]. Preclinical research also suggests a possible synergistic mechanism with lithium, involving activation of the mTOR pathway and inhibition of GSK-3 [189]. Reported side effects were generally moderate, and the overall incidence of treatment-emergent mania or hypomania was low [190], although small sample sizes limit the generalizability of these findings.

Despite promising rapid-onset effects, the long-term safety profile of ketamine and psychedelic compounds remains a subject of concern. With repeated ketamine use, studies have reported risks of dissociation, urinary tract dysfunction (e.g., cystitis), memory impairment, and potential neurotoxicity, especially at higher doses or with chronic exposure [191]. In BD specifically, ketamine use requires caution due to risk of mood switching or induction of mania, particularly with repeated administration [192]. Psychedelics such as psilocybin and ayahuasca also present unresolved questions around perceptual disturbances, psychological destabilization, and contraindications in patients with psychotic-spectrum features. Therefore, current clinical trials are carefully excluding individuals at risk, and long-term, placebo-controlled studies are necessary to assess safety, tolerability, and recurrence risk.

Basically, investigating classic psychedelics in patients with BD remains controversial, as case reports have documented manic switches following psychedelic use. Safety remains a central concern in the use of rapid-acting treatments such as ketamine. Patients with BD often present with comorbid vulnerabilities, such as substance use disorders, compulsive gaming or hypersexual behaviors, which may be exacerbated by interventions that acutely elevate mood or impulsivity. Of particular concern is the possibility of suicidality rebound following the transient relief provided by ketamine: individuals may experience a brief improvement followed by rapid emotional decline if the therapeutic effect dissipates abruptly. These risks underscore the importance of comprehensive behavioral and psychiatric screening, patient education, and close post-treatment monitoring. Ethical considerations, including informed consent regarding transient benefits and safety planning for high-risk populations, should be integrated into treatment protocols. Therefore, careful screening and monitoring of manic symptoms should be standard practice in psychedelic research involving clinical populations. Further longitudinal studies with larger, independent samples are essential to establish the safety and efficacy of psychedelic therapies in mood disorders [190].

4.6. Epigenetic Drugs

There is increasing evidence that epigenetic dysregulation, such as altered DNA methylation, histone modification, and microRNA expression, may play a role in the pathophysiology of BD. This has led to the exploration of so-called ‘epigenetic drugs’. However, the term warrants clarification: in this context, it refers primarily to agents that modulate gene expression via defined epigenetic mechanisms (e.g., histone deacetylase inhibitors). It should be noted that some drugs exhibit epigenetic effects without having been designed for that purpose, and it remains difficult to determine whether epigenetic changes observed in postmortem or peripheral tissue studies are due to medication, illness progression, or environmental confounders. At present, the application of epigenetic knowledge to treatment remains experimental, and translation into clinical practice is still in early stages.

Abnormalities in DNA methylation, histone modifications, and microRNA expression have been observed both in postmortem brain tissues and peripheral cells of individuals with BD. One notable finding is the overexpression of DNA methyltransferase 1 (DNMT1) in GABAergic neurons in the prefrontal cortex, leading to hypermethylation and reduced expression of key genes such as reelin and GAD67, both of which are crucial for synaptic function and inhibitory neurotransmission. These abnormalities support the rationale for targeting epigenetic enzymes to reverse transcriptional repression and restore normal neuronal activity [193].

Epigenetic drug research in BD is still largely at the preclinical stage, with a focus on histone deacetylase inhibitors (HDACi). Several HDACi, including valproic acid (already in clinical use), sodium butyrate, and trichostatin A, have demonstrated efficacy in animal and cellular models of BD. These agents have been shown to increase histone acetylation, enhance expression of BDNF, GAD67, and other neuroprotective genes, and mitigate mania-like behaviors such as hyperlocomotion. For example, valproic acid has been found to activate promoter IV of the BDNF gene in rat cortical neurons and regulate glutamate carboxypeptidase II stability in astrocytes. Sodium butyrate has been shown to reduce oubain-induced hyperactivity and improve oxidative stress parameters in rodent models of mania [193]. It should be stressed that these molecular findings should not be conflated with clinical efficacy, which remains to be established in well-designed trials.

The therapeutic promise of epigenetic drugs in BD lies in their potential to target the underlying molecular architecture of the illness, which may not be addressed adequately by traditional mood stabilizers. While valproic acid already exerts epigenetic effects as part of its mood-stabilizing profile, ongoing efforts aim to develop more selective HDAC2 inhibitors with cognitive-enhancing properties. Despite these promising findings, clinical translation remains limited, and further studies are needed to evaluate efficacy, safety, and long-term outcomes.

4.7. Chronotherapeutic Strategies in BD

Chronotherapy refers to interventions that align treatment with individual biological rhythms, including circadian and ultradian cycles. While lithium has long been known to influence circadian timing, growing evidence suggests that other pharmacological agents, such as SSRIs, atypical antipsychotics, and melatonergic compounds (e.g., agomelatine, ramelteon), also possess phase-shifting properties. These effects may support more biologically congruent dosing schedules. In BD, dysregulation of cortisol and melatonin rhythmicity is well-documented, particularly in manic and depressive phases. Individualized chronobiological profiling may help optimize treatment timing. A compelling neuroscientific explanation comes from animal models showing dopaminergic ultradian (~4-h) rhythms associated with mania-like symptoms, findings that are supported by human neuroimaging data linking wake onset regularity to positive affect and mood instability in BD [194]. Furthermore, melatonin, while discussed in mitochondrial contexts, may be more effectively framed as a core chronotherapeutic agent, owing to its role in circadian entrainment. Behavioral regularity, including consistent sleep and meal timing, also offers non-pharmacological avenues to stabilize disrupted rhythms, which in turn may reduce mood lability and immune dysregulation.

Despite being a relatively young field, chronotherapy represents a promising integrative approach that aligns with the shift toward precision and personalized psychiatry.

5. Digital Health and Future Treatment Monitoring

5.1. Wearable Devices and Symptom Monitoring, Relapse Prevention

Wearable devices such as smartwatches and biometric sensors enable the real-time collection of physiological data, monitoring parameters including heart rate, heart rate variability (HRV), sleep quality, and physical activity. These data can be used to detect changes in physiological patterns that precede manic or depressive episodes, allowing for timely intervention [195]. In traditional clinical practice, mood and symptom monitoring is primarily based on subjective self-assessments and periodic consultations. Wearable devices, by contrast, provide objective and continuous data, which are useful for detecting changes in sleep (e.g., reduction of REM sleep before manic episodes), monitoring declines in physical activity associated with depressive states, and tracking heart rate as a marker of physiological arousal [196].

Thus, the advantage over traditional practice is clear, as wearable devices provide a continuous, rather than episodic, picture of the patient’s state, filling the informational gap between clinical visits. Longitudinal data analysis enables the identification of recurring patterns preceding crises. For instance, reduced sleep duration can be an early warning sign of a manic shift, whereas increased sedentary behavior may indicate an impending depressive episode [196]. Early recognition of such signals allows for preemptive interventions, potentially preventing symptom worsening.

Furthermore, patients who actively monitor their own data (e.g., through apps connected to their smartwatch) tend to develop greater awareness of their behavioral and emotional states, adhere more closely to pharmacological prescriptions, and report changes in a timely manner. This improved engagement significantly enhances the clinician–patient relationship [195]. Effective monitoring reduces the likelihood of severe relapses and potential hospitalizations, benefiting both patient quality of life and healthcare system costs [197].

5.2. Digital Applications for Monitoring Bipolar Disorder (BD)

Psychiatric apps (e.g., ERPOnline, MONARCA, MindLAMP, EmoTrack, etc.) use a combination of active and passive data collection to anticipate relapse symptoms and initiate targeted interventions before a full-blown crisis occurs [198,199,200].

Active monitoring requires the patient’s participation, involving daily self-reporting of perceived mood (via visual scales or instruments like PANAS, Positive and Negative Affect Schedule, a widely used self-report questionnaire designed to measure two primary dimensions of mood), sleep quality and duration, energy levels, activity, perceived stress, significant life events, and adherence to medication. Users complete brief questionnaires or respond to daily prompts. These responses are recorded and analyzed to detect significant changes relative to personalized thresholds [201].

Passive monitoring involves the automatic collection of data through smartphone sensors or connected devices. These data include movement and GPS-based location (with reduced mobility possibly indicating social withdrawal or depression), phone usage patterns (reduced communication as a depressive marker, increased communication as a manic marker), sleep analysis via accelerometry or smartwatches, and speech or linguistic analysis (changes in speech rate, tone, or content may reflect manic logorrhea or depressive slowing) [202]. These data are processed in real time by algorithms that recognize patterns associated with prodromal manic or depressive states. Apps may send automatic alerts to the patient or notify clinicians [202].

5.3. Personalized Feedback and Interventions

Once an increased risk is detected, apps can suggest corrective actions: breathing exercises, grounding techniques, increased physical activity; encourage contact with healthcare providers; dynamically adjust medication reminders or mood diary prompts; and issue clinical alerts if a predefined risk threshold is exceeded.

A concrete example is ERPOnline (Enhanced Relapse Prevention Online), which is based on cognitive-behavioral therapy (CBT) for bipolar disorder. It includes interactive modules that help patients identify personalized early warning signs, relapse management sheets, and practical suggestions [203]. ERPOnline has been shown to enhance patient awareness of personal patterns and reduce relapse severity [202,204].

5.4. Sensors for Lithium Monitoring: Emerging Technologies

5.4.1. Wearable and Continuous-Readout Biosensors

These devices aim to detect lithium levels in real time using alternative body fluids such as sweat, saliva, or interstitial fluid. Electrochemical sensors (wearables), utilizing electrodes modified with lithium-selective materials (e.g., graphene, conductive polymers), can be applied to the skin via adhesive patches or integrated into smartwatches. They detect lithium in sweat and convert chemical signals into electrical outputs [205].

5.4.2. Saliva- or Oral Fluid-Based Sensors

Previous studies have shown a strong correlation between lithium concentrations in saliva and serum, validating saliva as an alternative biological fluid for therapeutic monitoring [206]. A multicenter study analyzing 169 paired saliva-serum samples from 75 patients confirmed that saliva can provide reliable estimates of serum lithium levels. Predictive accuracy was improved by incorporating clinical covariates such as daily dose, administration schedule, smoking status, and diabetes. Using intra-subject longitudinal data (e.g., saliva/serum ratios averaged over three previous visits) yielded high accuracy, especially in younger patients. In older individuals, greater variability was observed, possibly due to decreased renal function and polypharmacy [207]. Saliva collection offers several advantages over blood draws: it is non-invasive, well tolerated (even during manic episodes or inpatients), does not require trained personnel, and can be performed at home. Its operational flexibility supports more frequent monitoring and more responsive, personalized care. Additionally, lithium’s stability in saliva at room temperature for at least 24 h, and its compatibility with postal shipping, make it ideal for telemonitoring programs.

Technologically, these findings support the development of saliva-based point-of-care (POC) devices for lithium measurement. Such devices could integrate miniaturized sensors and wireless data transmission systems (via smartphone or cloud), providing a practical and accessible solution for continuous or on-demand monitoring. Unlike current POC serum tests, which are expensive and complex due to blood component separation, salivary tests require minimal processing and are easier to implement. Nevertheless, standardization of collection procedures and further investigation into potential confounding factors (e.g., diet, hydration, caffeine intake, circadian rhythm) are needed to ensure method reliability.

In summary, saliva represents a highly promising fluid for lithium monitoring. The development of dedicated saliva-based sensors could revolutionize bipolar disorder management by making treatment safer, more personalized, and more accessible. Currently, lithium treatment requires frequent monitoring (e.g., blood tests every 1–3 months) and careful management of toxicity and side effects (renal, thyroid, cognitive). Smart sensors could enable dynamic dosage adjustments based on real-time body levels, prevent toxicity or underdosing without waiting for clinical deterioration, reduce monitoring-related anxiety, and improve adherence—a critical issue given the current decline in lithium use for practical reasons.

5.5. The Role of Artificial Intelligence in BD Management and Precision Psychiatry

Artificial intelligence (AI) can analyze large volumes of heterogeneous, longitudinal data (e.g., mood, sleep, activity, speech, biometric data) collected via wearables, mobile apps, and sensors. These seemingly trivial data may contain subtle patterns predictive of relapse. Relevant technologies include: Machine Learning, for models that learn from individual historical data to forecast future episodes; Recurrent Neural Networks (RNNs) or LSTM networks, particularly effective for time-series data; Natural Language Processing (NLP), to analyze patients’ language in digital diaries or conversations, capturing early signs (e.g., manic or depressive tone) [208,209,210].

AI can integrate various data types: clinical data (diagnosis, treatment history), genomic data (pharmacogenomics, genetic variants affecting drug response), digital data (from apps, wearables, social media); biomarkers (inflammation, neuroimaging, electroencephalography). This allows the construction of predictive, personalized models for each patient to recommend the most effective pharmacological treatment, forecast side effects or poor response, and optimize dosages in real time (e.g., lithium or mood stabilizers) [208,209].

AI may act as a digital co-pilot for psychiatrists, offering automatic data synthesis, predictive relapse risk visualization, evidence-based treatment suggestions (clinical decision support systems, CDSS), and patient stratification by risk or phenotype. The clinician remains the final decision-maker, but AI can reduce cognitive burden and error risk, enhancing the precision and timeliness of clinical decisions [209].

6. Integration with Psychotherapy and Lifestyle Interventions

The management of BD is anticipated to incorporate new individualized psychotherapeutic approaches, lifestyle modifications, and advanced pharmacotherapy techniques. While pharmacotherapy is a crucial component, it tends to be insufficient by itself in achieving enduring remission, particularly during the interepisodic phases. These periods, where residual symptoms pose notable psychosocial functioning and life quality challenges, can significantly hinder psychosocial functioning and quality of life [211]. A summary of types of interventions, clinical objectives, and neurobiological targets are offered in

Table 1. Psychotherapy and lifestyle interventions, Clinical Objectives, and Neurobiological Targets for BD.

Intervention	Clinical Objective	Neurobiological Target
Cognitive Behavioral Therapy (CBT)	Reduce depressive/manic symptoms, improve adherence and cognitive regulation	Dorsolateral prefrontal cortex, amygdala
Interpersonal and Social Rhythm Therapy (IPSRT)	Stabilize circadian and social rhythms, prevent relapses	Suprachiasmatic nucleus, HPA axis
Psychoeducation	Increase illness awareness, improve compliance	Cognitive modulation, indirect neuroplasticity
Family-Focused Therapy (FFT)	Reduce family stress and negative emotional expression	Ventromedial prefrontal cortex, limbic circuits
CBT for Insomnia (CBT-I)	Improve sleep quality, reduce mood instability	Medial prefrontal cortex, HPA axis regulation
Nutritional Interventions	Reduce inflammation, improve neurochemical balance	Gut-brain axis, microglia, cytokines
Aerobic Exercise	Enhance mood, cognition, and neuroplasticity	BDNF, hippocampus, systemic inflammation reduction
Digital Monitoring Tools	Personalize treatment, improve symptom tracking	Functional regulation via biofeedback and behavioral data

.

6.1. Personalized Psychotherapy and Neurobiological Targets

More advanced psychotherapy modalities, such as CBT, Interpersonal and Social Rhythm Therapy (IPSRT), Psychoeducation, and Family therapy (FFT) have demonstrated positive impacts on the management of BD. Tailored CBT for BD enhances medication compliance by diagnosing unhelpful beliefs and reducing depressive and manic relapses [212].

IPSRT has been proven to defend the stability of social and circadian rhythms, both of which are commonly disrupted in individuals with BD. The technique is centered on the regularity of bodily and social interactions, the principal neurobiological targets of which include the suprachiasmatic nucleus and the hypothalamic-pituitary-adrenal (HPA) axis. Multiple studies provide evidence for its efficacy: Moot et al. (2022) [213] showed there were changes in social and occupational functioning; Orhan et al. (2024) [214] validated its feasibility and acceptability in group settings; a systematic review by Aktaş and Dülgerler (2024) [215] also verified its efficacy for sustaining euthymia and psychosocial stability.

Neuroimaging studies have further advanced understanding of the mechanisms of therapeutic effects from such interventions. Goldstein-Piekarski et al. found neural circuit biotypes corresponding to transdiagnostic symptom dimensions, such as anhedonia, anxious avoidance, or emotional dysregulation. Their results point to the critical role of prefrontal-limbic and salience network connectivity, indicating that treatments such as CBT or IPSRT may have beneficial effects through the alteration of these neural systems [216]. Evidence from randomized controlled trials supports the efficacy of structured psychotherapies such as CBT, IPSRT, psychoeducation, and FFT. Tailored CBT protocols for BD have demonstrated efficacy in reducing depressive and manic symptomatology, improving medication adherence, and modifying dysfunctional thought patterns [212]. Orhan et al. (2024) found that group-format IPSRT was both feasible and effective in older adults, yielding significant improvements in quality of life [214].

6.2. Lifestyle Interventions: Sleep, Diet, and Exercise

Both sleep and physical activity have been overlooked in psychosocial treatment, yet they are vital for long-term prognosis and clinical stability in BD. Mood episodes can be triggered by sleep disturbances, which can also occur as a symptom. CBT for Insomnia (CBT-I) tailored for BD prompts stable affective states and continuity during sleep. In a clinical trial by Harvey et al. (2015), participants receiving CBT-I enhanced sleep efficiency and decreased mood fluctuations [217].

Dietary interventions are also gaining attention. Anti-inflammatory and neuroprotective nutritional strategies, such as increased intake of omega-3 fatty acids, polyphenols, and B vitamins, are linked to symptom reduction and modulation of immune-inflammatory responses. A systematic review [218] confirmed significant improvements in depressive symptoms through dietary modification. Moreover, emerging research on the gut-brain axis indicates that microbiome composition may influence neuroinflammation and emotional regulation. Several studies highlighted the potential role of gut microbiota, and probiotics and prebiotics in particular, in enhancing mood stability [219,220,221].

Structured physical exercise, especially aerobics, is highly beneficial to one’s neurobiological resilience. Ana Sylvia and colleagues found relations in patients where engagement in exercise was linked to higher BDNF levels, better functioning of the hippocampal region, lowered inflammation, and thus improved mood and cognition [222]. Improved mood and enhanced sleep patterns have been described with sustained interventions. This included sleep therapy tailored for chronic insomnia, even within the framework of BD [223]. Nutritional psychiatry remains a useful complement with the strongest symptoms of inflammation and immune response depression arising from certain diets having anti-inflammatory and neuroprotective properties rich in omega-3, polyphenol, and B vitamins [224].

Research regarding gut-brain axis has correlated diversity within microbiota to neuroinflammation and modulation of mood making them good candidates for probiotic and prebiotic therapies [219,221,225]. Structured aerobic therapies have provided results such as higher BDNF levels, lower inflammation, and enhanced mood [226,227].

6.3. Toward Integrated and Personalized Care

The integration of pharmacological, psychotherapeutic, and lifestyle-based strategies marks the next frontier in personalized treatment for bipolar disorder. An ideal care model incorporates next-generation agents, such as glutamatergic modulators, mitochondrial enhancers, and epigenetic compounds, alongside evidence-based therapies like CBT, IPSRT, and FFT. These multi-layered interventions target both symptom reduction and functional recovery. Digital health technologies have become key allies in this paradigm. Smartphone applications, wearable devices, and machine learning algorithms now enable real-time monitoring of sleep, activity, and mood, allowing clinicians to adapt interventions dynamically. According to Torous et al. (2021) [195], these tools improve engagement, enable early detection of relapse, and enhance individualization of care. Ultimately, this multidimensional, biopsychosocial framework supports the goals of precision psychiatry, shifting the focus from symptom control to functional recovery and quality of life, with interventions adapted to each individual’s biological, psychological, and environmental profile. This biopsychosocial framework aligns with the emerging vision of precision psychiatry: a shift from symptom control to functional recovery, with interventions customized to the individual’s biological, psychological, and social context.

7. Conclusions and Future Directions

As our understanding of BD continues to expand, the field of pharmacotherapy is undergoing a revolution. Older treatment paradigms, centered largely on mood stabilizers, antipsychotics, and adjunctive antidepressants, are giving way to a more intricate and patient-tailored approach. This is part of a larger movement within psychiatry toward precision medicine, in which treatment is increasingly guided by the integration of biological, behavioral, and environmental data. As this review has established, BD is now not solely thought of in terms of neurotransmitter imbalance but rather as a heterogeneous neuropsychiatric disorder with genetic, inflammatory, mitochondrial, and circadian dysregulation. This new paradigm not only changes our conceptualization of BD but also opens up new therapeutic possibilities. Precision medicine will likely guide future care models in BD; however, it must be grounded in precision research. This entails studying biologically and clinically well-characterized individuals, rather than relying on heterogeneous populations, and integrating longitudinal trajectories of illness. An understanding of the dynamic etiopathology of BD, including alternating phases (mania, depression, mixed states) and progressive staging, should inform both treatment development and individualized therapeutic planning.

Most promising is the convergence of a number of fields, such as neuroscience, genomics, bioinformatics, and digital health, that will shape future care models. The future of pharmacological therapy will probably be dominated by multimodal strategies in the use of traditional psychopharmacological treatments and newer agents such as glutamatergic modulators, neurosteroids, mitochondrial enhancers, and epigenetic agents. The potential therapeutic benefits of drugs such as ketamine, zuranolone, and psilocybin in bringing about rapid change in treatment-resistant symptoms are tempered by concerns regarding their safety profiles and long-term effects. Nevertheless, their rapid action and capacity to address core symptoms such as suicidality and anhedonia represent a paradigm shift in the treatment of bipolar depression. Similarly, the modulation of neuroinflammation with agents like minocycline, pioglitazone, and omega-3 fatty acids has shown promising, although preliminary, results. While results remain heterogeneous, they allude to the role of immunomodulation in the clinical trajectory of BD.

The integration of digital technology into psychiatric practice adds another dimension to this development. Continuous monitoring of physiological and behavioral patterns using wearables and smartphone apps enables pre-emptive detection of relapse signatures, real-time therapy modifications, and patient engagement. Importantly, these tools also offer a means of collecting ecologically valid data in naturalistic settings, which can be exploited through artificial intelligence to tailor treatment algorithms and forecast disease course. In particular, AI-powered systems have the potential to assist clinicians in the future by selecting the most optimal pharmacologic treatments based on individual biomarker profiles, adherence patterns, and risk predictions, taking psychiatry one step closer to the goals of personalized medicine.

There are significant challenges to cope with. Most of the compounds discussed here await robust, large-scale clinical verification. Long-term treatment effect durability, affective switching risks, and long-term safety profiles of novel agents remain to be thoroughly investigated. Heterogeneity in study designs, patient populations, and outcome measures further undermines the comparability and generalizability of results. There is also an urgent requirement for robust biomarkers to guide diagnosis, forecast response to therapy, and monitor disease progression. Future research should prioritize longitudinal, multimodal investigations that integrate genetic, neuroimaging, immunological, and digital phenotyping information. Clinical trials must go beyond symptom reduction as an endpoint. To align with precision medicine principles, future studies should incorporate biologically stratified inclusion criteria, based on specific biomarkers or chronobiological phenotypes, rather than broad diagnostic categories. Moreover, outcome measures should include cognitive function, psychosocial integration, biological target modulation (e.g., immune or mitochondrial markers), and quality of life. This shift from symptom-centric to mechanism-based trial design represents a necessary evolution in BD research. Lastly, a transdiagnostic approach, recognizing the shared neurobiological substrates of mood, anxiety, and psychotic disorders, may offer more inclusive therapeutic targets and treatment strategies. To truly support the development of precision medicine, clinical studies should incorporate inclusion and exclusion criteria based on the biological phenotypes or biomarkers the intervention is designed to modulate, such as inflammatory cytokines, circadian rhythm profiles, or mitochondrial function, rather than relying solely on symptom severity or diagnostic categories. This would enable more targeted, hypothesis-driven trials with greater translational relevance.

BD management over the next few years will probably depend on the effective integration of various modalities: pharmacological innovation based on biological plausibility, digital technologies that take care outside the clinic, and psychotherapeutic and lifestyle interventions matched to the individual’s needs and abilities. Collaborative research partnerships, patient-centered models of care, and flexible policy frameworks will be crucial to the translation of scientific advances into tangible improvements in clinical outcomes.