Bipolar disorder and somatic diseases: Focus on oxidative stress markers Authors: Sinem Balaç, İzel Cemre Aksahin, Hidayet Ece Arat-Çelik, Şermin Genc, Deniz Ceylan Published: 2026 ABSTRACT Bipolar disorder (BD) is associated with a substantial burden of somatic diseases, including cardiovascular, neurodegenerative, and inflammatory conditions. However, the biological continuity linking the core pathophysiology of BD to these somatic manifestations remains poorly understood. Oxidative stress is increasingly recognized as a shared biological process across many medical conditions that commonly co-occur with BD and has been extensively investigated using diverse biomarkers, consistently demonstrating redox imbalance across illness stages and mood states. This evidence supports oxidative stress as a core biological process contributing to multisystem vulnerability rather than merely a secondary consequence of affective episodes.This review employed a narrative synthesis approach. Somatic diseases most frequently observed in BD were identified through systematic reviews and meta-analyses, and oxidative stress–related biomarkers associated with BD were examined alongside their alterations in commonly comorbid somatic diseases. Evidence from epidemiological syntheses, biomarker studies, and mechanistic research was integrated to identify converging findings, inconsistencies, and key research gaps. Across systematic syntheses, thirteen major non-communicable somatic disease categories were consistently associated with BD, particularly cardiometabolic, cardiovascular, endocrine, inflammatory, and neurodegenerative conditions. In parallel, BD research demonstrates alterations across multiple oxidative stress domains, with the most consistent findings observed for lipid peroxidation markers (e.g., TBARS, MDA) and oxidative nucleic acid damage (e.g., 8-oxo-dG, 8-OHGuo), whereas evidence for protein oxidation and antioxidant enzyme activity is more heterogeneous and state dependent. Notably, no studies directly examine associations between oxidative stress markers and objectively measured somatic comorbidity or multimorbidity in BD, and current evidence therefore relies on indirect triangulation between robust redox alterations in BD and oxidative stress mechanisms established in general medical populations.Overall, oxidative stress appears to represent a systemic biological feature of BD that aligns mechanistically with the somatic diseases contributing to its mortality gap. However, biomarker heterogeneity, predominantly cross-sectional study designs, and the lack of integrated somatic phenotyping limit clinical interpretation. Longitudinal, state-sensitive studies linking redox dysregulation with somatic outcomes are needed to determine whether oxidative stress reflects cumulative systemic vulnerability rather than episodic illness activity. Introduction Bipolar disorder (BD) is a lifelong psychiatric condition characterized by recurrent manic and depressive episodes and substantial disability. Beyond impairments in vocational, social, and financial functioning, BD is associated with markedly reduced life expectancy, driven not only by elevated suicide risk but also by a high burden of somatic morbidity (Singh et al., 2025a). This somatic burden reflects two closely related domains: somatic symptoms intrinsic to the disorder, such as fatigue, sleep disturbances, pain, and appetite changes, and somatic diseases, including cardiometabolic, inflammatory, and neurodegenerative diseases. Together, interacting behavioral, treatment-related, healthcare, genetic, and biological factors both arise from and generate a cumulative somatic burden that worsens illness course, treatment response, relapse risk, functioning, quality of life, and healthcare utilization, with links to potentially shared mechanisms such as mitochondrial dysfunction, inflammation, and oxidative stress (Özerdem et al., 2022). Despite its clear clinical relevance, the biological processes linking somatic burden to illness progression in BD remain insufficiently theorized. Individuals with BD consistently show elevated rates of obesity, diabetes, cardiovascular disease, renal disease, and immune-related conditions, with approximately 34–41% of individuals with BD living with three or more medical conditions, most commonly cardiometabolic and endocrine disorders. This multimorbidity substantially increases healthcare utilization and is associated with an approximately twofold increase in premature mortality compared with the general population (Weiner et al., 2011; Westman et al., 2013; SayuriYamagata et al., 2017; Biazus et al., 2023). Furthermore, extensive systematic reviews indicate that, while the cumulative effect is particularly pronounced in older adults, somatic health issues in BD are not confined to specific age groups (Dragasek et al., 2023). This highlights the need for research into the underlying causes that go beyond just diagnosing BD. We need integrated models that consider the overall vulnerability of the system (Warner et al., 2023; Almeida et al., 2016). Bridging somatic disease burden and BD biology therefore requires mechanistic models that integrate affective illness trajectories with multisystem biological processes. However, prevailing clinical and nosological frameworks continue to rely on a dichotomous view of bodily manifestations, in which physiological alterations without an immediately identifiable organic lesion are often relegated to residual or nonspecific categories, effectively labeling unexplained bodily phenomena rather than interrogating their underlying biology. Although somatic disease comorbidity occurs at broadly comparable rates and with similar clinical impact in depression and BD, its biological underpinnings have been far more extensively studied in depression, where immunometabolic subtypes have been incorporated into contemporary diagnostic and research frameworks and linked, in large cohort studies such as the Netherlands Study of Depression and Anxiety (NESDA), to low-grade inflammation and metabolic dysregulation (Kapfhammer, 2006; Lamers et al., 2013; Lamers et al., 2020; Milaneschi et al., 2020). In contrast, the biological architecture underlying somatic disease vulnerability in BD remains poorly characterized, underscoring a substantial gap in mechanistic understanding. The allostatic load model posits that the biological toll of BD stems from the combined effects of recurring mood episodes and chronic stress, resulting in enduring vulnerability across multiple somatic systems (Kapczinski et al., 2008). Oxidative stress, in this context, serves as a key mechanism connecting the course of mood disorders with an elevated risk of somatic illness (Özerdem et al., 2022). Meta-analyses have repeatedly shown disrupted redox homeostasis and oxidative damage to lipids, proteins, and nucleic acids in individuals with BD, with these abnormalities persisting even during periods of stable mood (Andreazza et al., 2008; Brown et al., 2014; Jiménez-Fernández et al., 2021), and transdiagnostic evidence further situates BD within a broader redox-related vulnerability to cardiometabolic and neurodegenerative disease (Jorgensen et al., 2022; Arat-Çelik et al., 2024; Çeli et al., 2023). In this narrative review, we examine oxidative stress as a central biological process relevant to somatic disease burden in BD, with brief consideration of regulatory components of redox homeostasis. We integrate evidence from systematic reviews, meta-analyses, and mechanistic studies to identify conceptual and methodological gaps, highlight inconsistencies across the literature, and delineate key priorities for future research. Importantly, rather than reiterating established epidemiological evidence on medical comorbidity, we focus on whether oxidative stress markers capture biologically meaningful processes relevant to systemic disease vulnerability in BD. Methods This review was conducted using a structured narrative synthesis approach to integrate clinical and biological evidence relevant to somatic disease burden, oxidative stress biology, and BD. Literature searches were performed primarily in PubMed/MEDLINE, with supplementary screening of reference lists from key systematic reviews and meta-analyses. Only human studies published in English were considered. Studies were included if they met the following criteria: (i) systematic reviews or meta-analyses examining somatic comorbidity or physical health outcomes in bipolar disorder; (ii) studies investigating oxidative stress–related biomarkers in bipolar disorder; (iii) systematic evidence examining oxidative stress pathways in somatic diseases associated with BD; (iv) studies reporting biomarkers reflecting key domains of oxidative stress in somatic diseases; (v) human studies published in English. To identify somatic diseases most consistently associated with BD, an initial focused PubMed search was conducted using combined title and abstract–based queries including bipolar disorder and terms related to somatic comorbidity (e.g., physical, somatic, medical, bodily).) Search focused on noncommunicable diseases, and infectious diseases were excluded. Search results were restricted to systematic reviews and meta-analyses and limited to human studies published in English. This focused strategy was intended to identify high-level evidence syntheses explicitly addressing somatic comorbidity in BD and to anchor the review in robust epidemiological findings. This search was complemented by broader title/abstract searches and backward and forward citation tracking of key systematic reviews to enhance the comprehensiveness of the literature identification process. For Table 1, a structured and domain-specific literature search was conducted in PubMed to identify evidence on oxidative stress–related biomarkers in BD, with separate searches performed for each molecular damage domain to ensure conceptual clarity. Searches targeted lipid peroxidation, protein oxidation, nucleic acid damage, and antioxidant defense pathways using combinations of bipolar disorder and oxidative stress–related keywords within title and abstract fields. The literature was first screened for studies jointly assessing oxidative stress–related biomarkers in BD and somatic diseases. In the absence of such evidence, an indirect synthesis approach was adopted by integrating BD biomarker findings with systematic evidence from somatic disease research. Across all oxidative stress searches, results were screened to identify systematic reviews and meta-analyses synthesizing human data. Evidence was integrated at the biomarker-domain level to summarize biological matrices, measurement approaches, mechanistic interpretation, and the robustness and consistency of findings in BD. This approach was designed to synthesize the most clinically interpretable and translationally relevant evidence into an integrated conceptual framework of somatic disease vulnerability in BD rather than to conduct a formal systematic review.Table 1Integrated evidence from meta-analyses on oxidative damage domains, antioxidant responses associations in BD.Table 1Damage DomainMarkerSpecimen & Quantification TechniquesWhat It ReflectsFinding in BDKey ReferencesLipid damageTBARS, MDAPredominantly peripheral blood–derived matrices (serum/plasma). Quantified as MDA or TBARS using spectrophotometric assays; substantial methodological overlap across studies.Systemic lipid oxidative injury, membrane vulnerability, mitochondrial ROS generation, impaired antioxidant buffering.↑ Consistently elevated vs. healthy controls. TBARS ↑ across mood states (mania, depression, euthymia); MDA shows greater heterogeneity. Both state- and trait-related components reported.(Andreazza et al., 2008; Brown et al., 2014; Jiménez-Fernández et al., 2021; Capuzzi et al., 2022; Almulla et al., 2023)Protein damageProtein carbonyls, 3-nitrotyrosine (3-NT)Peripheral blood samples assessed using DNPH-based protein carbonyl assays, ELISA, or immunochemical detection of 3-NT.Oxidative/nitrosative modification of amino acid side chains, altered protein structure/function, proteasomal stress.↔ No robust pooled increase at meta-analytic level; high heterogeneity. Some evidence for ↑ 3-NT in subsets → likely state-dependent.Brown et al. (2014)Nucleic acid damage8-OHdG/8-oxo-dG, 8-oxoGuo, DNA/RNA strand breaksPeripheral (urine, plasma, serum, leukocytes) and postmortem brain tissue. Quantified via LC-MS/MS, ELISA; strand breaks via Comet assay.Direct oxidative DNA/RNA injury; genomic instability; transcriptional and epigenetic disruption.↑ Robustly elevated across peripheral and central matrices. One of the largest and most consistent effect sizes among oxidative domains; both state- and trait-like properties reported.(Brown et al., 2014; Arat-Çelik et al., 2024; Kucuker et al., 2022; Black et al., 2015)Indirect oxidative stress indicatorsSOD, CAT, GPx, GSH/GSSG; NRF2/KEAP1 signalingPeripheral blood; spectrophotometric or kinetic enzyme assays; Peripheral blood transcriptomics; limited postmortem brain protein data.Global antioxidant defense and compensatory vs. exhausted redox responses; Master regulator of antioxidant and cytoprotective gene expression.↔/↑/↓ Heterogeneous. Direction depends on mood state, illness stage, and medication; no robust pooled effect; ↓ Reduced NRF2/KEAP1 expression reported in BD; findings limited and heterogeneous; clinical relevance unclear.(Jiménez-Fernández et al., 2021; Arat-Çelik et al., 2024; Black et al., 2015; Lima et al., 2022; Zhang et al., 2018; García-Álvarez et al., 2018; García et al., 2019)Glutathione systemGSH, GSH/GSSG ratioPeripheral blood; colorimetric or fluorometric assaysCellular redox buffering capacity; vulnerability to oxidative overload.↓ Reduced GSH across mood states; partial normalization in remission.(Brown et al., 2014; Jiménez-Fernández et al., 2021; Ballatori et al., 2009)Abbreviations: 8-OH-dG, 8-hydroxy-2′-deoxyguanosine; 8-oxo-dG, 8-oxo-7,8-dihydro-2′-deoxyguanosine; APE1, apurinic/apyrimidinic endonuclease 1; BER, base excision repair; CAT, catalase; CSF, cerebrospinal fluid; GSH, glutathione; GPx, glutathione peroxidase; MDA, malondialdehyde. Table 1Integrated evidence from meta-analyses on oxidative damage domains, antioxidant responses associations in BD.Table 1Damage DomainMarkerSpecimen & Quantification TechniquesWhat It ReflectsFinding in BDKey ReferencesLipid damageTBARS, MDAPredominantly peripheral blood–derived matrices (serum/plasma). Quantified as MDA or TBARS using spectrophotometric assays; substantial methodological overlap across studies.Systemic lipid oxidative injury, membrane vulnerability, mitochondrial ROS generation, impaired antioxidant buffering.↑ Consistently elevated vs. healthy controls. TBARS ↑ across mood states (mania, depression, euthymia); MDA shows greater heterogeneity. Both state- and trait-related components reported.(Andreazza et al., 2008; Brown et al., 2014; Jiménez-Fernández et al., 2021; Capuzzi et al., 2022; Almulla et al., 2023)Protein damageProtein carbonyls, 3-nitrotyrosine (3-NT)Peripheral blood samples assessed using DNPH-based protein carbonyl assays, ELISA, or immunochemical detection of 3-NT.Oxidative/nitrosative modification of amino acid side chains, altered protein structure/function, proteasomal stress.↔ No robust pooled increase at meta-analytic level; high heterogeneity. Some evidence for ↑ 3-NT in subsets → likely state-dependent.Brown et al. (2014)Nucleic acid damage8-OHdG/8-oxo-dG, 8-oxoGuo, DNA/RNA strand breaksPeripheral (urine, plasma, serum, leukocytes) and postmortem brain tissue. Quantified via LC-MS/MS, ELISA; strand breaks via Comet assay.Direct oxidative DNA/RNA injury; genomic instability; transcriptional and epigenetic disruption.↑ Robustly elevated across peripheral and central matrices. One of the largest and most consistent effect sizes among oxidative domains; both state- and trait-like properties reported.(Brown et al., 2014; Arat-Çelik et al., 2024; Kucuker et al., 2022; Black et al., 2015)Indirect oxidative stress indicatorsSOD, CAT, GPx, GSH/GSSG; NRF2/KEAP1 signalingPeripheral blood; spectrophotometric or kinetic enzyme assays; Peripheral blood transcriptomics; limited postmortem brain protein data.Global antioxidant defense and compensatory vs. exhausted redox responses; Master regulator of antioxidant and cytoprotective gene expression.↔/↑/↓ Heterogeneous. Direction depends on mood state, illness stage, and medication; no robust pooled effect; ↓ Reduced NRF2/KEAP1 expression reported in BD; findings limited and heterogeneous; clinical relevance unclear.(Jiménez-Fernández et al., 2021; Arat-Çelik et al., 2024; Black et al., 2015; Lima et al., 2022; Zhang et al., 2018; García-Álvarez et al., 2018; García et al., 2019)Glutathione systemGSH, GSH/GSSG ratioPeripheral blood; colorimetric or fluorometric assaysCellular redox buffering capacity; vulnerability to oxidative overload.↓ Reduced GSH across mood states; partial normalization in remission.(Brown et al., 2014; Jiménez-Fernández et al., 2021; Ballatori et al., 2009)Abbreviations: 8-OH-dG, 8-hydroxy-2′-deoxyguanosine; 8-oxo-dG, 8-oxo-7,8-dihydro-2′-deoxyguanosine; APE1, apurinic/apyrimidinic endonuclease 1; BER, base excision repair; CAT, catalase; CSF, cerebrospinal fluid; GSH, glutathione; GPx, glutathione peroxidase; MDA, malondialdehyde. Integrated evidence from meta-analyses on oxidative damage domains, antioxidant responses associations in BD. Abbreviations: 8-OH-dG, 8-hydroxy-2′-deoxyguanosine; 8-oxo-dG, 8-oxo-7,8-dihydro-2′-deoxyguanosine; APE1, apurinic/apyrimidinic endonuclease 1; BER, base excision repair; CAT, catalase; CSF, cerebrospinal fluid; GSH, glutathione; GPx, glutathione peroxidase; MDA, malondialdehyde. Results Most frequent somatic diseases in BD Using a title- and abstract–based search strategy targeting BD and related physical health outcomes, two evidence syntheses addressing somatic comorbidity in BD were identified: an umbrella review of meta-analyses (Kang et al., 2024) and a systematic review focusing on older adults with BD (Warner et al., 2023). The umbrella review by Kang et al. included 12 meta-analyses, which themselves included 145 primary studies and over 60 million participants from 29 countries (Kang et al., 2024). This review looked at 14 different physical health outcomes. It found a significant link between BD and eight noncommunicable diseases. There was strong evidence for hypertension, some evidence for dementia and Parkinson's disease, and weak to moderate evidence for obesity, type 2 diabetes, asthma, and breast cancer. Cardiometabolic disorders showed the most consistent associations, even though there was considerable variability in the outcomes. Warner and colleagues' systematic review also found that older adults with BD had a higher incidence of noncommunicable somatic comorbidities. Cardiovascular disease was the most common, with an average prevalence exceeding 30%. Diabetes, cancer, renal disease, and obesity were also more prevalent. However, the review's conclusions were limited by the small number of studies, the lack of matched control groups, and differences in how outcomes were defined. Consistent with the current study, communicable diseases, such as toxoplasmosis, were excluded, and only noncommunicable diseases were included in the following analyses. Analyses were restricted to noncommunicable somatic conditions. Accordingly, results were reported for diabetes mellitus, hypertension and cardiovascular disease, obesity, asthma, migraine, dementia and Parkinson's disease, hypothyroidism, multiple sclerosis, renal disease and breast cancer (Fig. 1).Fig. 1Physical health comorbidities in bipolar disorder with shared oxidative stress and DNA damage pathways. Physical health comorbidities in bipolar disorder with shared oxidative stress and DNA damage pathways. Oxidative stress biomarkers in BD research Across studies, multiple oxidative stress markers measured in blood, urine, and CSF indicate increased oxidative damage in BD, involving lipids, proteins, and nucleic acids, together with alterations in antioxidant defenses and DNA repair pathways, as reflected by increased lipid peroxidation (malondialdehyde, MDA; thiobarbituric acid reactive substances, TBARS), oxidative DNA damage (8-oxo-2′-deoxyguanosine, 8-oxo-dG), altered DNA repair capacity (8-oxoguanine DNA glycosylase, OGG1; apurinic/apyrimidinic endonuclease 1, APE1), reduced antioxidant reserve (glutathione, GSH), altered enzymatic antioxidant activity (catalase, CAT; glutathione S-transferase, GST; superoxide dismutase, SOD; glutathione peroxidase, GPX), and dysregulation of master antioxidant signaling (NRF2, Keap1). Lipid Peroxidation: Markers of lipid peroxidation, particularly malondialdehyde (MDA) and thiobarbituric acid–reactive substances (TBARS), were consistently elevated in individuals with BD compared with healthy controls across multiple meta-analyses (Brown et al., 2014; Jiménez-Fernández et al., 2021). The first meta-analysis including 13 studies reported increased TBARS levels with a large effect size (Andreazza et al., 2008), which was confirmed by a subsequent meta-analysis of 27 studies identifying lipid peroxidation markers as those with the highest effect sizes (Brown et al., 2014). More recent meta-analyses similarly demonstrated higher MDA and TBARS levels in BD (Jiménez-Fernández et al., 2021). TBARS levels were increased across illness phases, with higher levels reported during manic episodes compared with euthymia, suggesting predominantly state-related effects (Jiménez-Fernández et al., 2021). Findings for MDA were more heterogeneous, with elevations reported during mood episodes and euthymia, including unmedicated individuals (Capuzzi et al., 2022), in contrast to other meta-analytical findings (Jiménez-Fernández et al., 2021). With respect to illness stage, TBARS findings varied: while some studies found no differences between early- and late-stage BD (Grewal et al., 2023; Panizzutti et al., 2015), others reported increased TBARS levels during depressive episodes and euthymia in late-stage BD, with more restricted increases during depressive episodes in early-stage BD (Siwek et al., 2016), or elevated levels in both stages compared with healthy controls (Kunz et al., 2008). Protein Oxidation and Nitrosative Stress: Meta-analytical evidence did not demonstrate significant differences in protein carbonyl content (PCC) or 3-nitrotyrosine (3-NT) levels between individuals with BD and healthy controls (Brown et al., 2014). Consistent with this, no differences in PCC levels were observed between early- and late-stage BD (Panizzutti et al., 2015; Andreazza et al., 2009). In contrast, increased 3-NT levels were reported in both early- and late-stage BD in one study (39). Nitrosative stress markers showed more consistent alterations. Nitric oxide (NO) (Brown et al., 2014; Savas et al., 2006) and total nitrite levels (Jiménez-Fernández et al., 2021) were elevated in BD; and a positive correlation was observed between NO levels and the number of manic episodes (Savas et al., 2006), supporting an association between illness burden and nitrosative stress (Singh et al., 2025b). Nucleic Acid Damage: Markers of oxidative DNA and RNA damage, including 8-hydroxy-2′-deoxyguanosine (8-OH-dG) and 8-hydroxyguanosine (8-OHGuo), were significantly elevated in BD across serum, blood, urine, and CSF samples (Brown et al., 2014; Çeli et al., 2023; Knorr et al., 2019; Goh et al., 2021). Postmortem studies further demonstrated increased DNA fragmentation in brain tissue from individuals with BD (Benes et al., 2003; Mustak et al., 2010). RNA damage markers showed both state- and trait-like properties. Longitudinal CSF studies demonstrated persistently elevated 8-oxoGuo levels across mood states and follow-up, with further increases during new episodes (Knorr et al., 2019). In contrast, DNA damage markers appeared predominantly state-dependent, with elevations most consistently observed during depressive episodes (Çeli et al., 2023). During euthymia, findings for 8-OH-dG were inconsistent across studies (Ceylan et al., 2018a, 2018b; Huzayyin et al., 2014; Jacoby et al., 2016). Illness burden was positively correlated with 8-OH-dG levels (Soeiro-de-Souza et al., 2013). DNA Repair Pathways: Studies focusing on the base excision repair (BER) pathway demonstrated altered expression of DNA repair genes in BD. Reduced expression of OGG1 and POLG was reported in BD (Munkholm et al., 2015), with independent replication confirming decreased OGG1 expression (Ceylan et al., 2018a). Longitudinal data showed increased 8-OH-dG levels and reduced OGG1 expression during depressive episodes, followed by partial normalization during remission (Ceylan et al., 2020). In contrast, increased extracellular OGG1 protein levels were reported in one study (Ceylan et al., 2019), whereas more recent findings showed no significant difference in OGG1 mRNA expression but reduced NEIL1 expression in depressed individuals with BD compared with healthy controls (Mart et al., 2022). Indirect Oxidative Stress Indicators: Meta-analytical pooling of 44 studies demonstrated altered antioxidant profiles in BD, including lower glutathione levels and higher uric acid concentrations (Jiménez-Fernández et al., 2021). Antioxidant enzyme activities, particularly catalase (CAT) and glutathione S-transferase (GST), were increased, whereas findings for superoxide dismutase (SOD) and glutathione peroxidase (GPX) were inconsistent (Jiménez-Fernández et al., 2021). Across illness stages, no consistent differences were observed in SOD (Lima et al., 2022) or GPX activity (Lima et al., 2022; Panizzutti et al., 2015; Andreazza et al., 2009) between early- and late-stage BD; while elevated GST and glutathione reductase activity were reported predominantly in late-stage BD (Andreazza et al., 2009). Uric acid levels were higher during manic episodes compared with depressive episodes (Jiménez-Fernández et al., 2021), and catalase activity showed a negative correlation with illness duration (Selek et al., 2015). Limited evidence is available regarding alterations in the NRF2 signaling pathway in bipolar disorder and related psychiatric conditions. Additionally, to date, three studies have directly examined components of the NRF2 pathway in human samples. In a longitudinal study of patients with first-episode mania, García et al. (2019) measured NRF2 DNA-binding transcriptional activity in nuclear extracts from peripheral blood mononuclear cells (PBMCs) as well as in plasma samples. Among 60 patients assessed at baseline and 44 completing follow-up at six months, NRF2 activity increased over time, with a more pronounced trend observed in non-adherent patients. A postmortem investigation by Zhang et al. (2018) assessed NRF2 and KEAP1 protein levels in the parietal cortex (Brodmann area 7) of individuals with major depressive disorder, bipolar disorder, and schizophrenia (n = 15 per group) compared with healthy controls (n = 15). Both NRF2 and KEAP1 protein levels were significantly reduced in all psychiatric groups relative to controls. Similarly, García-Álvarez et al. (2018) measured nuclear NRF2 protein levels in PBMCs from individuals with schizophrenia (n = 123), bipolar disorder (n = 102), and healthy controls (n = 80). NRF2 levels showed an upward trend in the schizophrenia group relative to bipolar disorder and control groups, although the difference did not reach statistical significance. Collectively, these findings suggest that empirical evidence on NRF2 signaling in bipolar disorder remains limited and heterogeneous, with studies differing in sample type, illness stage, and methodological approach. Finally, our research showed the strongest and most consistent cross-disease evidence observed for lipid peroxidation, oxidative DNA damage, and glutathione-antioxidant system impairment. Somatic diseases and BD associated oxidative stress markers Despite extensive investigation of oxidative stress–related biomarkers in BD, no studies were identified that directly examined oxidative stress or oxidative DNA damage markers in relation to objectively defined somatic diseases or multimorbidity within BD cohorts. Across the oxidative stress literature, individuals with bipolar disorder and clinically significant medical comorbidities are frequently excluded, or somatic disease status is insufficiently characterized, precluding integrated analyses. As a result, the existing evidence base consists of parallel but disconnected research streams: one documenting robust alterations in redox biology in BD, and another establishing strong associations between oxidative stress pathways and somatic diseases in the general biomedical literature. This structural gap necessitates an indirect integrative approach to evaluate the potential relevance of oxidative stress markers to somatic disease burden in BD.s. In mood disorders, oxidative stress has been proposed as convergent pathways linking central nervous system dysfunction to peripheral somatic pathology via immune–inflammatory activation, endothelial dysfunction, and metabolic dysregulation (Berk et al., 2011; Moylan et al., 2014). In BD, meta-analysis provides consistent support for redox abnormalities. Andreazza and colleagues concluded that oxidative stress markers are increased in BD and positioned oxidative stress as a plausible contributor to BD biology (Andreazza et al., 2008). Brown and colleagues also reinforced this conclusion and emphasized oxidative damage to DNA/RNA and lipids as prominent features of BD-associated oxidative stress signatures (Brown et al., 2014). Jiménez-Fernández and colleagues further demonstrated that oxidative stress parameters and antioxidant defenses vary by mood polarity and euthymic status, supporting the presence of both state-linked and persisting alterations and highlighting the risk of misclassification when mood state and illness stage are not modeled explicitly (Jiménez-Fernández et al., 2021). A critical interpretive issue is heterogeneity of biomarkers and biological matrices. Meta-analyses aggregate lipid peroxidation measures (e.g., MDA/TBARS), antioxidant enzymes, glutathione-system readouts, and nucleic acid oxidation products, each reflecting partially distinct compartments and processes (Andreazza et al., 2008; Brown et al., 2014; Jiménez-Fernández et al., 2021). This heterogeneity strengthens the rationale for pathway-level and multimarker approaches, particularly when the clinical target is somatic burden rather than a narrow symptom state. Associations between oxidative stress pathways and somatic diseases Despite extensive investigation of oxidative stress–related markers in BD, no studies have directly examined oxidative stress or oxidative DNA damage markers in relation to objectively defined somatic diseases or multimorbidity within BD cohorts. Individuals with clinically significant medical comorbidities are frequently excluded, or somatic disease status is insufficiently characterized, precluding integrated analyses (Table 1). Fig. 2 illustrates the current evidence showing the predefined somatic diseases associated with BD and shared oxidative stress and DNA damage pathways, including diabetes mellitus, hypertension and cardiovascular disease, obesity, asthma, migraine, dementia and Parkinson's disease, hypothyroidism, multiple sclerosis, renal disease, and breast cancer (Fig. 2; Supplementary Table).Fig. 2Converging oxidative stress pathways linking bipolar disorder to somatic burden. Converging oxidative stress pathways linking bipolar disorder to somatic burden. Across disease categories, alterations were consistently observed in oxidative DNA base damage, DNA strand instability, lipid peroxidation (Daiber et al., 2021, Filipek et al., 2025, Furukawa et al., 2004, Hecht et al., 2016, Santibáñez et al., 2025), antioxidant defense systems, glutathione availability, and base excision repair (BER)–related pathways. Markers of oxidative DNA base damage, most consistently 8-oxo-dG/8-OH-dG, were elevated across cardiometabolic, neurological, endocrine, renal, and oncological conditions. In BD, increased 8-oxo-dG has been reported in peripheral blood, urine, cerebrospinal fluid, and postmortem brain tissue, with variation across illness states and episodes (Andreazza et al., 2008; Brown et al., 2014; Jiménez-Fernández et al., 2021). In the somatic literature, meta-analyses and systematic reviews reported elevated 8-OH-dG in diabetes, hypertension/cardiovascular disease, obesity, and neurodegenerative disorders, whereas evidence for asthma, hypothyroidism, renal disease, multiple sclerosis, and breast cancer was mainly derived from primary studies and broader oxidative stress syntheses (Graille et al., 2020; Hazukova et al., 2024; Di Minno et al., 2016; Schrag et al., 2013; Piñar-Morales et al., 2025; Vodošek Hojs et al., 2020; Qing et al., 2019; Delic et al., 2025). DNA strand breaks and fragmentation, most commonly assessed using comet-based assays, were consistently increased in BD and were associated with illness duration and episode burden (Andreazza et al., 2008; Brown et al., 2014). In somatic conditions, increased DNA strand damage was reported in diabetes, obesity, asthma, renal disease, and hypothyroidism, while pooled meta-analytic evidence was available only for hypertension (Hazukova et al., 2024; Sardaş et al., 2001; Kasznicki et al., 2009; Usman et al., 2018; Zeyrek et al., 2009; Schupp et al., 2016). For migraine, dementia, Parkinson's disease, multiple sclerosis, and breast cancer, evidence was largely descriptive. Markers of lipid peroxidation, including malondialdehyde (MDA) and TBARS, showed the most consistent pattern across conditions. In BD, lipid peroxidation markers were elevated, particularly during acute mood episodes (Andreazza et al., 2008; Brown et al., 2014; Jiménez-Fernández et al., 2021)). In parallel, systematic reviews and meta-analyses reported increased MDA/TBARS levels across all examined somatic conditions, including diabetes, cardiovascular disease, obesity, asthma, migraine, neurodegenerative disorders, hypothyroidism, multiple sclerosis, renal disease, and breast cancer (Schrag et al., 2013; Lodovici et al., 2011; Marseglia et al., 2015; Sultana et al., 2013; Pasupuleti et al., 2020; Cammisotto et al., 2021; ERDAL et al., 2005; Sener et al., 2007; Vaziri, 2004; Neri et al., 2015). Across conditions, antioxidant enzyme systems (SOD, CAT, GPx) were dysregulated. In BD, overall reduced or functionally inadequate antioxidant activity was reported (Andreazza et al., 2008; Brown et al., 2014). Similar alterations were described in somatic diseases, including diabetes, cardiovascular disease, obesity, migraine, hypothyroidism, multiple sclerosis, renal disease, and breast cancer (Vaziri, 2004; Maritim et al., 2003; Förstermann et al., 2017; Vincent et al., 2006; Mancini et al., 2016; Gupta et al., 2012). The glutathione system was also consistently altered. Reduced GSH levels were reported across mood states in BD, with partial normalization during remission (Brown et al., 2014; Jiménez-Fernández et al., 2021). In the somatic literature, systematic reviews documented glutathione depletion across all examined disease categories (Schrag et al., 2013; Maritim et al., 2003; Mancini et al., 2016; Rahman et al., 2006; Heitzer et al., 2001; Uttara et al., 2009; Aoyama et al., 2013). Finally, alterations in BER-related pathways were reported. In BD, reduced expression of OGG1 and APE1 was observed in brain and peripheral tissues, alongside familial vulnerability signals (Çeli et al., 2023; Ceylan et al., 2020). In somatic diseases, BER involvement was supported by mechanistic and genetic association studies, particularly in cardiovascular disease, neurodegeneration, renal disease, and breast cancer; however, disease-specific systematic reviews or meta-analyses pooling BER biomarkers were not identified, except for cancer-focused genetic meta-analyses of BER polymorphisms (Zhao et al., 2022; Wu et al., 2022; Liu et al., 2020; Qiao et al., 2018). Discussion Current evidence indicates that individuals with BD carry a substantially increased burden of somatic morbidity, with cardiometabolic conditions such as diabetes mellitus and hypertension occurring at markedly higher rates than in the general population, alongside consistent alterations in oxidative stress pathways, among which markers of DNA damage represent some of the most robust and stable findings. Despite this consistency, the present findings do not provide direct evidence that oxidative stress constitutes a primary causal mechanism underlying somatic vulnerability in BD, largely due to the absence of studies jointly examining somatic comorbidity and oxidative stress markers within the same cohorts. However, based on indirect evidence, our data synthesis indicates that oxidative stress markers altered in BD are similarly disrupted in BD-associated somatic conditions, suggesting that oxidative stress may reflect a shared biological mechanism underlying both psychiatric and somatic vulnerability. In addition to oxidative damage markers, alterations in antioxidant defense systems provide important insight into the biological context of redox dysregulation in BD. The GSH system represents the principal intracellular antioxidant buffer, protecting cellular structures from reactive oxygen species and maintaining redox balance. Reduced GSH levels reported across mood states in BD may therefore reflect diminished cellular resilience to oxidative stress rather than transient metabolic fluctuations (Brown et al., 2014; Jiménez-Fernández et al., 2021). Similarly, enzymatic antioxidant defenses such as SOD, CAT, and GPx show heterogeneous alterations across illness phases. These findings may represent compensatory responses to increased oxidative load in early disease stages, followed by functional exhaustion in chronic illness (Jiménez-Fernández et al., 2021; Lima et al., 2022; Andreazza et al., 2009). Comparable patterns of glutathione depletion and antioxidant enzyme dysregulation are widely reported in cardiometabolic and inflammatory disorders, supporting the hypothesis that impaired antioxidant defense systems may contribute to multisystem vulnerability rather than representing a disorder-specific feature of BD. Oxidative stress provides a unifying perspective for elucidating the somatic manifestations of BD by perturbing interrelated systems, including mitochondrial bioenergetics, inflammatory signaling pathways, endothelial function, and nociceptive processing; this, in turn, establishes a connection between affective disorders and pain, fatigue, and the comorbid patterns of cardiometabolic and inflammatory conditions (SayuriYamagata et al., 2017; Berk et al., 2011; Halliwell et al., 2015). Furthermore, across chronic inflammatory, metabolic, vascular, and age-related ailments, oxidative stress markers consistently present a clinical signature, with oxidative DNA damage (8-oxo-dG/8-OHdG) serving as an indicator of cumulative systemic burden, rather than reflecting disorder-specific effects (Khalil et al., 2012; Valavanidis et al., 2009; Urbaniak et al., 2020). The clinical consequences of such damage are contingent upon the efficacy of DNA repair mechanisms, specifically base excision repair; its dysfunction is associated with accelerated aging, an elevated risk of cancer, and increased susceptibility to chronic diseases (Gorini et al., 2021). These pathways, in concert, offer a plausible biological foundation for the heightened cardiovascular and metabolic risks observed in BD, thereby supporting redox imbalance as a potential link between affective disorders and long-term somatic vulnerability, rather than a mechanism specific to the disorder (Weiner et al., 2011; Westman et al., 2013; SayuriYamagata et al., 2017). Nevertheless, key regulatory mechanisms such as the NRF2 pathway remain underexplored, and future studies should employ more comprehensive, pathway-level biomarkers (Morgenstern et al., 2024; Bhandari et al., 2021). Oxidative stress responses are tightly regulated by the NRF2 signaling pathway, a central transcriptional regulator of cellular antioxidant and cytoprotective responses. Under conditions of oxidative stress, NRF2 dissociates from its cytoplasmic inhibitor KEAP1 and translocates to the nucleus, where it activates antioxidant response element (ARE)–dependent genes involved in glutathione synthesis, detoxification enzymes, and mitochondrial protection. Beyond its role in redox homeostasis, NRF2 signaling also exerts anti-inflammatory and cytoprotective effects and has been implicated in the pathophysiology of multiple systemic disorders, including cardiometabolic and neurodegenerative diseases (Brandes et al., 2020). Although preliminary evidence suggests altered NRF2/KEAP1 signaling in BD, human data remain limited and heterogeneous. From a translational perspective, pharmacological activation of NRF2 has emerged as a potential strategy for modulating oxidative stress pathways. NRF2 activators have been explored in several experimental models of oxidative stress–related diseases and may provide useful tools for hypothesis testing in preclinical models of mood disorders. Future studies examining NRF2 signaling in BD—particularly in relation to somatic disease burden—may therefore help clarify whether redox dysregulation represents a mechanistic driver of systemic vulnerability or a downstream consequence of illness processes. Despite robust and consistent evidence of oxidative stress dysregulation in BD, these observations have yet to be effectively incorporated into clinically relevant or outcome-oriented models. A primary constraint is the enduring gap between biological indicators and clinical manifestations. Limited research directly investigates the correlation between oxidative stress, oxidative DNA damage, or repair mechanisms and somatic symptom domains, including pain and fatigue, or prospectively predicts cardiometabolic and vascular outcomes that contribute to increased mortality and multimorbidity in BD (Westman et al., 2013; Biazus et al., 2023). While oxidative DNA damage and base excision repair offer plausible mechanistic connections between persistent redox imbalance, somatic illness, and accelerated aging, these pathways remain inadequately examined within BD research. Furthermore, the existing literature is constrained by fragmented single-marker methodologies, heterogeneous assay techniques, and a predominance of cross-sectional study designs. Biological hypotheses in BD are, in a broader sense, inadequately correlated with the most significant clinical outcomes. Although substantial epidemiological data indicate that cardiometabolic disease and cardiovascular mortality are the principal factors contributing to diminished life expectancy (Weiner et al., 2011; Westman et al., 2013; SayuriYamagata et al., 2017), research on oxidative stress is largely divorced from the prediction of outcomes. Furthermore, trait vulnerability is infrequently differentiated from episode-related biological fluctuations, and the temporal relationships between redox dysregulation, mitochondrial dysfunction, and the onset of somatic disease remain largely untested. Consequently, while oxidative stress is recognized as a systemic characteristic of BD, its clinical significance is not definitively established. The persistent use of case-control comparisons, without comprehensive somatic phenotyping, extended molecular endpoints, and outcome-focused methodologies, has impeded translational advancements. Substantial progress necessitates integrative, state-sensitive, and longitudinal models that directly connect redox dysregulation, genomic damage, and DNA repair mechanisms to somatic burden and cardiometabolic risk, the factors that ultimately determine morbidity and mortality in BD. Finally, BD is associated with a markedly increased burden of somatic comorbidity, particularly cardiometabolic conditions, alongside consistent alterations in oxidative stress pathways, with oxidative DNA damage representing one of the most robust biological findings. However, current evidence does not directly support oxidative stress as a primary causal mechanism of somatic vulnerability in BD, largely because somatic outcomes and redox markers have rarely been examined within the same cohorts. Indirect evidence, however, suggests that the oxidative stress markers altered in BD are similarly disrupted in somatic conditions associated with BD. This implies a shared biological basis for both psychiatric and somatic vulnerabilities, rather than a process unique to a specific disorder. To clarify the clinical importance of these pathways, state-sensitive, longitudinal studies are needed. These studies should combine detailed somatic phenotyping with assessments of redox imbalance, genomic damage (e.g., 8-oxo-dG/8-OHdG), DNA repair capacity, and key regulatory pathways such as NRF2, and directly relate these processes to cardiometabolic risk and long-term outcomes in BD. Conflict of interest Authors have no conflict of interest.