Prevalence and risk factors of bipolar disorder in central nervous system autoimmune demyelinating diseases Authors: Omid Mirmosayyeb, Sepehr Aghajanian, Saeed Vaheb, Fateme Mohammadifard, Mohammad Reza Sharbafchi, Vahid Shaygannejad Published: 2026 ABSTRACT AbstractBackgroundAffective disorders, including bipolar disorder (BPD), are increasingly recognized in patients with multiple sclerosis (pwMS), yet their comparable prevalence across MS subtypes and in neuromyelitis optica spectrum disorder (NMOSD) remains unclear. This study aimed to estimate the prevalence of BPD among patients with MS and NMOSD and to identify clinical and demographic factors associated with increased BPD risk.MethodsWe conducted a cross-sectional study involving 4,460 patients with MS and 299 patients with NMOSD. Psychiatric assessments were performed using structured interviews based on DSM-IV criteria. Demographic, neurological, and treatment data were collected from medical records and clinician evaluations. Multivariate logistic regression was used to assess factors associated with BPD, including disease subtype, disability level (EDSS), and clinical presentation.ResultsThe overall prevalence of BPD was 3.07%, with the highest rate in SPMS (5.63%), followed by NMOSD (4.68%). The risk of BPD was comparable among NMOSD vs. MS patients (OR:1.15; 95%CI:0.60–2.17). Among pwMS, RRMS (2.48%) and PPMS (1.42%) had statistically comparable rates of BPD (OR: 0.55; 95%CI: 0.17–1.76); however, SPMS patients had highest prevalence of BPD that was statistically greater than RRMS (OR: 2.32; 95%CI: 1.57–3.44). Higher EDSS, single relationship status, addiction, smoking, and cerebellar symptoms at onset were also associated with higher likelihood of BPD.ConclusionsHigher disability in SPMS, early cerebellar involvement, and psychosocial factors were associated with higher odds of BPD. These findings underscore the need for routine psychiatric screening and tailored support strategies in high-risk subgroups.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12888-026-08310-8.Keywords: Bipolar disorder, Neuromyelitis optica, Multiple sclerosis, Affective disorder, Aquaporin-4, Epidemiology Introduction Multiple sclerosis (MS) is a chronic inflammatory disorder of the central nervous system (CNS), characterized by demyelination in the brain and spinal cord, which results in a wide range of sensory and motor impairments [1]. In addition to its physical symptoms, MS is frequently accompanied by psychiatric and mood disorders, which contribute significantly to overall disease burden [2, 3]. Depression is the most commonly reported psychological comorbidity and has been linked to poorer treatment adherence, reduced functional capacity, lower quality of life, and increased suicide risk. Other psychiatric conditions associated with MS include bipolar disorder (BPD), anxiety disorders, psychosis, substance use disorders, and personality disorders [4, 5]. Several studies have consistently shown that depression, anxiety, and BPD are more prevalent in individuals with MS compared to the general population [2, 3, 6]. Variable factors, including the neuroinflammatory nature of MS plaques, lesion location, family history of psychiatric disorders, pharmacological side effects, and psychosocial stressors, can elevate the risk of psychiatric comorbidities in MS patients [7, 8]. Although a large Swedish cohort reported a higher prevalence of BPD among MS patients, among individuals with MS, this association was not significantly explained by familial factors, suggesting a potential disease-related mechanism [9]. Lifetime prevalence estimates for BPD in the MS population vary widely, ranging from 0% to 16.2% across studies [3], highlighting methodological differences and the heterogeneous nature of the disease. Potential contributors to this increased vulnerability could be mitochondrial downregulation, neuroinflammation, MS lesions in specific areas, and lower volume of the putamen and nucleus accumbens [10, 11]. Still, the precise mechanisms underlying the elevated risk of BPD in individuals with MS remain under investigation. Neuromyelitis Optica Spectrum Disorder (NMOSD), another chronic inflammatory and demyelinating condition, predominantly affects the optic nerves, spinal cord, and brainstem, often presenting with more severe attacks than MS. Although NMOSD was historically considered to spare the brain, newer imaging and pathological studies have shown brain involvement in a substantial subset of patients, including regions implicated in cognition and emotion. Psychological and psychiatric manifestations—including depression, anxiety, and fatigue—are commonly reported in NMOSD and may be comparable to MS [12–14]. However, the presence and impact of BPD in NMOSD remain largely understudied. To date, few if any large-scale investigations have systematically evaluated BPD prevalence in NMOSD cohorts, underscoring a critical gap in our understanding of psychiatric comorbidities in this population. In view of the gap regarding the psychological burden of MS subtypes and NMOSD with regards to BPD, we aimed to estimate the prevalence of BPD across different subtypes of MS and in NMOSD, and investigate other risk factors that may contribute to a higher burden of BPD in patients who suffer from these central nervous system autoimmune demyelinating diseases (CNS ADD). Methods This cross-sectional study was conducted between January 2018 and December 2024 at the Department of Neurology at Kashani Hospital and the Hakim MS Clinic in Isfahan, Iran. Eligible participants included adult patients diagnosed with MS or NMOSD. MS diagnosis was established in accordance with the 2017 revision of the McDonald criteria [15] and aquaporin-4-positive NMOSD diagnoses followed the 2015 criteria issued by the International Panel for NMO Diagnosis (IPND-2015) [16]. Patients with clinically isolated syndrome (CIS), or radiologically isolated syndrome (RIS), myelin oligodendrocyte glycoprotein antibody-associated disease (MOGAD), transverse myelitis, acute disseminated encephalomyelitis (ADEM), and aquaporin-4 negative NMOSD cases were excluded. Aquaporin-4 antibodies were detected using enzyme-linked immunosorbent assay. To maximize sample size and capture the breadth of clinical presentations, participants were recruited using an availability-based, consecutive sampling approach during routine visits to the aforementioned centers. All included patients provided written informed consent after a full explanation of study protocol. The study was performed in accordance with the Declaration of Helsinki and its amendments and was approved by the Ethics Committee of Isfahan University of Medical Sciences (IR.ARI.MUI.REC.1401.061). Data collection Demographic, neurological, and psychiatric data were obtained through a combination of patient interviews, clinician-administered evaluations, and review of existing medical records. The diagnosis of bipolar disorder was assessed at the time of study enrollment through clinician-administered semi-structured interviews conducted by licensed board-certified psychiatrists. A diagnosis of BPD was made using Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) criteria and the Persian-translated Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I) diagnostic algorithms conducted by the aforementioned psychiatrists [17]. The choice of interview and diagnostic criteria was due to the fact that these represent validated and consistently implemented structured diagnostic instruments in our setting during the study period. Neurological assessments were performed by the expert neurologist affiliated with the MS specialty clinics. Diagnoses of comorbid conditions such as migraine or seizures were confirmed through clinical evaluation and chart review. The Expanded Disability Status Scale (EDSS) was assessed by trained providers during the most recent clinic visit and recorded from the medical records at baseline. Baseline data on the most prominent initial presenting symptom and disease-modifying therapies (DMTs) were also extracted from the patients’ medical records. Addiction was defined as a history of substance use disorder as documented in clinical records and/or structured interviews. Disease duration was defined from the time of definitive diagnosis to the date of enrollment. Information regarding medical comorbidities was collected during structured clinical interviews conducted at the time of enrollment. Active brain MRI plaques refer to gadolinium-enhancing brain lesions seen on the most recent MRI scan (within the last 6 months of inclusion). Statistical analysis We conducted summary descriptive analyses for both the total sample and the subset of patients with an MS diagnosis. Descriptive tables were provided for baseline demographics, disease specifications across all patients and patients with MS (pwMS) subgroup. Pearson’s chi-square and independent t-tests were used to compare patients with and without BPD as appropriate. Variables with a p-value < 0.20 in crude analyses and our pre-defined adjustment set were included in univariate and multivariate logistic regression models to explore potential risk modifiers influencing BPD prevalence between MS and NMOSD, as well as across MS subtypes. To address the disproportionate number of BPD cases to total subjects, we applied backward selection to reduce the number of predictors. Additionally, we performed adjusted analyses using a predefined minimal adjustment set for both comparisons. The minimal adjustment set included age, sex, marital status, disease course, disease duration, smoking, and DMTs. Evidence was prioritized for factors with statistical significance in the adjusted models. To examine differences in BPD prevalence across MS subtypes, we did not include both MS disease course and current EDSS scores in the same multivariable models to avoid redundancy and multicollinearity due to their strong interdependence. This was supported by an ANOVA test showing that EDSS scores differed significantly across MS subtypes (F(2, 4457) = 4769.4, p < 0.001), with disease course accounting for approximately 68% of the variance in EDSS scores (η² = 0.681). This strong correlation between disease course and disability, and the limited overlap in EDSS distributions across subtypes, precluded formal mediation analysis and models assessing subtype-specific risk were constructed without mutual adjustment for EDSS. All analyses were carried out in Stata version 18. Estimates with a p-value threshold of 0.05 were considered statistically significant. Results We screened a total of 5,131 individuals diagnosed with a variety of neurological disorders and 4,759 were included in the analyses after applying the inclusion criteria; after which, 4,460 were pwMS, including 3,502 with relapsing-remitting MS (RRMS), 212 with primary progressive MS (PPMS), and 746 with secondary progressive MS (SPMS). Additionally, 299 patients with NMOSD were included in the analysis (Fig. 1). The overall prevalence of BPD in the cohort was 3.07 per 100 patients. The crude prevalence was at 4.68 per 100 among NMOSD patients. Among MS subtypes, the prevalence was lowest in PPMS (1.42 per 100), intermediate in RRMS (2.48 per 100), and notably highest in SPMS (5.63 per 100), exceeding that observed in the NMOSD group. Flow diagram of the included participants Preliminary analyses identified several baseline variables associated with an increased prevalence of BPD among patients with CNS ADD. Married individuals were less likely to have BPD compared to those who were single (2.7% BPD rate vs. 4.3% BPD rate among single patients), and considering the limited sample size, none of the divorced patients were diagnosed with BPD. A history of addiction (15.8% among BPD patients vs. 7.8% in those without) and current smoking (17.8% vs. 8.8%) were also more common among patients with BPD. Patients with BPD had a longer disease duration at the time of evaluation compared to those without BPD (mean 13.1 vs. 11.7 years), although age at disease onset did not differ significantly between the groups (Table 1). Additionally, BPD was more frequently observed in patients who presented with certain initial symptoms and those with higher disability at the time of assessment, as measured by EDSS (mean 2.97 vs. 2.16). Table 1Comparison of clinical and demographic characteristics of MS and NMOSD patients with and without bipolar disorderVariableNumber availableBPD− (n = 4,613)BPD+ (n = 146)Total (n = 4,759)p-valueSex4,7590.198Female3,550 (77.0%)119 (81.5%)3,669 (77.1%)Male1,063 (23.0%)27 (18.5%)1,090 (22.9%)Age (Mean ± SD)4,75942.89 (10.27)43.40 (8.92)42.91 (10.23)0.559BMI (Mean ± SD)4,71424.51 (4.43)24.66 (4.85)24.51 (4.45)0.689Marital Status4,759 0.009 Divorced52 (1.1%)0 (0.0%)52 (1.1%)Married3,457 (74.9%)96 (65.8%)3,553 (74.7%)Single1,104 (23.9%)50 (34.2%)1,154 (24.2%)Education Level4,7590.143Illiterate (0)71 (1.5%)1 (0.7%)72 (1.5%)Diploma (12 yrs)1,736 (37.6%)69 (47.3%)1,805 (37.9%)Bachelor’s (16 yrs)1,603 (34.7%)42 (28.8%)1,645 (34.6%)Primary education (7 yrs)752 (16.3%)24 (16.4%)776 (16.3%)Master’s degree or higher (> 16)451 (9.8%)10 (6.8%)461 (9.7%)Disease classification4,759 < 0.001 NMOSD285 (6.2%)14 (9.6%)299 (6.3%)PPMS209 (4.5%)3 (2.0%)212 (4.4%)RRMS3,415 (74.0%)87 (59.6%)3,502 (73.6%)SPMS704 (15.3%)42 (28.8%)746 (15.7%)Disease duration4,75911.72 (6.57)13.12 (6.05)11.77 (6.56) 0.011 Age at onset4,75931.17 (9.59)30.28 (9.09)31.14 (9.58)0.269Primary onset symptom4,759 0.003 Brainstem441 (9.6%)19 (13.0%)460 (9.7%)Cerebellar211 (4.6%)16 (11.0%)227 (4.8%)Motor1,039 (22.5%)27 (18.5%)1,066 (22.4%)Visual1,290 (28.0%)40 (27.4%)1,330 (27.9%)Sensory1,632 (35.4%)44 (30.1%)1,676 (35.2%)EDSS at onset4,7592.281 (1.193)2.288 (1.037)2.282 (1.189)0.950Current EDSS4,7592.165 (1.772)2.969 (1.655)2.190 (1.773) < 0.001 Seizure4,759154 (3.3%)8 (5.5%)162 (3.4%)0.160Migraine4,759506 (11.0%)22 (15.1%)528 (11.1%)0.120Hyperlipidemia4,759242 (5.2%)6 (4.1%)248 (5.2%)0.543Diabetes4,759139 (3.0%)3 (2.1%)142 (3.0%)0.503Anemia4,759493 (10.7%)16 (11.0%)509 (10.7%)0.917Hyperthyroidism4,75951 (1.1%)1 (0.7%)52 (1.1%)0.630Hypothyroidism4,759484 (10.5%)15 (10.3%)499 (10.5%)0.933Addiction4,759359 (7.8%)23 (15.8%)382 (8.0%) < 0.001 Smoking4,759408 (8.8%)26 (17.8%)434 (9.1%) < 0.001 Active brain plaque4,759552 (12.0%)15 (10.3%)567 (11.9%)0.534 Comparison of clinical and demographic characteristics of MS and NMOSD patients with and without bipolar disorder Regarding treatment, there was no significant difference in the medications used by NMOSD patients with and without BPD. However, among patients with MS, treatment with monoclonal antibodies was more commonly associated with BPD (Table 2). Crude analyses suggested a trend toward a higher prevalence of BPD in NMOSD compared to MS (4.68% vs. 2.96%, p = 0.094), and a statistically significant difference among MS subtypes (p < 0.001), likely driven by the higher rate observed in SPMS (5.63%) relative to other MS subtypes (2.42%, p < 0.001). To account for potential confounders and to obtain more accurate estimates of associations with BPD, we conducted further analyses comparing NMOSD versus MS and among MS subtypes. Table 2Differences in disease phenotype and drug use in MS and NMOSD patientsVariableNumber availableBPD− (n = 4,613)BPD+ (n = 146)Total (n = 4,759)p-valueDisease classification (alternative classification)4,759 0.001 NMOSD285 (6.2%)14 (9.6%)299 (6.3%)RRMS3,415 (74.0%)87 (59.6%)3,502 (73.6%)PMS913 (19.8%)45 (30.8%)958 (20.1%)MS vs. NMO4,7590.094NMOSD285 (6.2%)14 (9.6%)299 (6.3%)MS4,328 (93.8%)132 (90.4%)4,460 (93.7%)MS subtypes4,460 < 0.001 PPMS209 (4.8%)3 (2.3%)212 (4.7%)0.174RRMS3,415 (78.9%)87 (65.9%)3,502 (78.5%) < 0.001 SPMS704 (16.3%)42 (31.8%)746 (16.7%) < 0.001 Progressive vs. Relapsing MS4,460 < 0.001 PMS913 (21.1%)45 (34.1%)958 (21.5%)RRMS3,415 (78.9%)87 (65.9%)3,502 (78.5%)Primary vs. secondary progressive MS958 0.010 PPMS209 (22.9%)3 (6.7%)212 (22.1%)SPMS704 (77.1%)42 (93.3%)746 (77.9%)Drugs (MS & NMOSD)4,759 0.028 Azathioprine31 (0.7%)3 (2.1%)34 (0.7%)Dimethyl Fumarate287 (6.2%)7 (4.8%)294 (6.2%)Fingolimod232 (5.0%)6 (4.1%)238 (5.0%)Glatiramer Acetate175 (3.8%)3 (2.1%)178 (3.7%)Interferons1,880 (40.8%)47 (32.2%)1,927 (40.5%)None58 (1.3%)0 (0.0%)58 (1.2%)Teriflunomide306 (6.6%)11 (7.5%)317 (6.7%)Monoclonal antibodies1,644 (35.6%)69 (47.3%)1,713 (36.0%)Drugs (categorized) 0.014 None58 (1.3%)0 (0.0%)58 (1.2%)Monoclonal1,644 (35.6%)69 (47.3%)1,713 (36.0%)Injection2,055 (44.5%)50 (34.2%)2,105 (44.2%)Oral856 (18.6%)27 (18.5%)883 (18.6%)Drugs (MS only)4,4600.092Dimethyl Fumarate287 (6.6%)7 (5.3%)294 (6.6%)Fingolimod232 (5.4%)6 (4.5%)238 (5.3%)Glatiramer Acetate175 (4.0%)3 (2.3%)178 (4.0%)Interferons1,880 (43.4%)47 (35.6%)1,927 (43.2%)None58 (1.3%)0 (0.0%)58 (1.3%)Teriflunomide306 (7.1%)11 (8.3%)317 (7.1%)Monoclonal antibodies1,390 (32.1%)58 (43.9%)1,448 (32.5%)Drugs (MS only, categorized)4,460 0.020 None58 (1.3%)0 (0.0%)58 (1.3%)Monoclonal antibodies1,390 (32.1%)58 (43.9%)1,448 (32.5%)Injection2,055 (47.5%)50 (37.9%)2,105 (47.2%)Oral825 (19.1%)24 (18.2%)849 (19.0%)Drugs (NMOSD patients only)2990.225Azathioprine31 (10.9%)3 (21.43%)34 (11.4%)Rituximab254 (89.1%)11 (78.6%)265 (88.6%) Differences in disease phenotype and drug use in MS and NMOSD patients Table 3 summarizes variables with a p-value < 0.2 in crude analyses that were subsequently included in the multivariable adjusted models. Notably, addiction (OR: 1.77; 95%CI: 1.03–3.03), higher EDSS scores, and cerebellar symptoms at onset (OR: 1.93; 95%CI: 1.10–3.39, compared to sensory onset) were significantly associated with increased odds of BPD and remained independent predictors after adjustment for other covariates including marital status, disease type, primary onset symptom, current EDSS, addiction, smoking, and Azathioprine use. Each one-point increase in EDSS was associated with a 30% higher likelihood of BPD (OR: 1.30; 95%CI: 1.20–1.42). Despite statistically similar prevalence rate of BPD in males (2.5%) vs. females (3.2%), male sex was less likely to be associated with BPD in MS & NMOSD after adjustment (OR: 0.42; 95%CI: 0.25–0.68). None of the DMTs, including monoclonal antibodies, were significantly associated with BPD after adjusting for potential confounders. Despite crude association of Azathioprine use with BPD, the drug was limited to NMOSD patients and did not show a significant association with BPD after adjustment, similarly monoclonal antibodies were administered in patients with advanced disease and higher EDSS and thus were no longer associated with BPD after adjustment for current disability. Overall, the adjusted comparison between NMOSD and MS revealed no statistically significant difference in BPD prevalence (OR: 1.15; 95%CI: 0.60–2.17). Table 3Multivariate analysis comparing the risk of BPD in both MS and NMOSDVariableUnivariate(Demographics + terms with P < 0.2)Minimal adjustment set(empirical)Backward selectionof significant terms(P = 0.1 threshold)OR (95% CI)p-valueOR (95% CI)p-valueOR (95% CI)p-value Age 1.00 (0.99–1.02)0.5590.99 (0.97–1.01)0.408-- Sex (Male vs. female) 0.76 (0.50–1.16)0.199 0.42 (0.25–0.68) 0.001 -- Marital status (Married vs. single) 0.61 (0.43–0.87) 0.006 0.59 (0.40–0.88) 0.010 0.61 (0.42–0.86) 0.006 Education level (vs. Diploma) Ref. Ref. Ref. Ref. Ref. Ref. Illiterate0.35 (0.05–2.59)0.306----Bachelor’s (16 yrs) 0.66 (0.45–0.97) 0.036 ----Primary education (7 yrs)0.80 (0.50–1.29)0.362----Master’s degree or higher (> 16)0.56 (0.28–1.09)0.088---- NMOSD vs. MS 1.61 (0.92–2.83)0.0981.35 (0.67–2.69)0.3991.15 (0.60–2.17)0.675 Disease duration (years) 1.03 (1.01–1.05) 0.012 1.02 (0.99–1.05)0.121-- Primary onset symptom (vs. Sensory) Ref. Ref. Ref. Ref. --Brainstem1.6 (0.92–2.76)0.094----Cerebellar 2.81 (1.56–5.07) 0.001 -- 1.93 (1.10–3.39) 0.021 Motor0.96 (0.59–1.57)0.882--0.67 (0.43–1.05)0.079Optic1.15 (0.74–1.78)0.528---- Current EDSS 1.22 (1.13–1.31) < 0.001 1.36 (1.21–1.53) < 0.001 1.30 (1.20–1.42) < 0.001 Seizure 1.68 (0.81–3.49)0.165---- Migraine 1.44 (0.91–2.29) 0.122 ---- Addiction 2.21 (1.40–3.50) 0.001 -- 1.77 (1.03–3.03) 0.039 Smoking 2.23 (1.44–3.45) < 0.001 3.3 (1.98–5.48) < 0.001 1.61 (0.96–2.69)0.069 Drugs (interferons) Ref. Ref. Ref. Ref. --Azathioprine 3.87 (1.14–13.11) 0.030 2.44 (0.58–10.17)0.2213.17 (0.82–12.23)0.094Dimethyl Fumarate0.98 (0.44–2.18)0.9520.93 (0.41–2.1)0.858--Fingolimod1.03 (0.44–2.45)0.9380.82 (0.34–1.97)0.664--Glatiramer Acetate0.69 (0.21–2.23)0.5300.8 (0.24–2.6)0.708--None------Teriflunomide1.44 (0.74–2.8)0.2861.46 (0.74–2.86)0.276--Monoclonal antibodies 1.68 (1.15–2.45) 0.007 0.76 (0.44–1.3)0.313-- Multivariate analysis comparing the risk of BPD in both MS and NMOSD Subgroup analysis on MS patients also suggested similar patterns to all patients regarding differences in MS subtypes, marital status, drug types, and addiction in those with BPD. Education status and lack of university education were also linked to higher BPD among pwMS. In addition to higher EDSS, being single, cerebellar symptoms at onset, and smoking were linked to higher BPD across all models on pwMS, roughly leading to a two-fold increase in the likelihood of BPD. In unadjusted analyses across MS subtypes (Table 4), patients with SPMS had significantly higher odds of BPD compared to those with RRMS (OR: 2.34; 95%CI: 1.61–3.41). This association remained robust after adjusting for relevant confounders (OR: 2.58; 95%CI: 1.36–4.92) and using backward selection (OR: 2.32; 95%CI: 1.57–3.44). In contrast, BPD odds did not differ significantly between PPMS and RRMS (OR: 0.55; 95%CI: 0.17–1.76), suggesting that the elevated risk was specific to the SPMS group. Furthermore, one-point increase in EDSS was associated with a higher likelihood of BPD in this subset (OR: 1.4, 95%CI: 1.23–1.60). Table 4Association between study variables and BPD in MS subgroupVariableUnivariate(Demographics + terms with P < 0.2)Minimal adjustment set with EDSS(empirical)Minimal adjustment set with MS subtypes(empirical)Backward selectionof significant terms excluding EDSS(P = 0.1 threshold)OR (95% CI)p-valueOR (95% CI)p-valueOR (95% CI)p-valueOR (95% CI)p-value Age 1.00 (0.99–1.02)0.5790.99 (0.97–1.01)0.3371 (0.98–1.02)0.816-- Sex (Male vs. female) 0.76 (0.48–1.18)0.224 0.4 (0.24–0.69) 0.001 0.47 (0.28–0.79)0.004-- Marital status (Married vs. single) 0.65 (0.45–0.94) 0.024 0.64 (0.42–0.98) 0.040 0.63 (0.42–0.96)0.0310.61 (0.41–0.89) 0.011 Education level (vs. Diploma) Ref. Ref. Ref. Ref. Ref. Ref. --Illiterate0.35 (0.05–2.59)0.306------Bachelor’s (16 yrs)0.66 (0.45–0.97) 0.036 ---- 0.64 (0.42–0.96) 0.033 Primary education (7 yrs)0.80 (0.50–1.29)0.362----0.64 (0.38–1.07)0.091Master’s degree or higher (> 16)0.56 (0.29–1.09)0.088---- 0.43 (0.19–0.96) 0.039 MS subtype (vs. RRMS) Ref. Ref. Ref. Ref. Ref. Ref. Ref. Ref. PPMS0.56 (0.18–1.80)0.332--0.70 (0.20–2.52)0.5900.55 (0.17–1.76)0.313SPMS 2.34 (1.61–3.41) < 0.001 --2.58 (1.36–4.92) 0.004 2.32 (1.57–3.44) < 0.001 Current EDSS 1.22 (1.13–1.32) < 0.001 1.4 (1.23–1.6) < 0.001 -- Omitted Omitted Disease duration (years) 1.03 (1.00–1.05) 0.020 1.02 (0.99–1.05)0.2631.02 (0.99–1.05)0.2461.02 (1.00–1.05)0.082 Primary onset symptom (vs. sensory) Ref. Ref. Ref. Ref. Ref. Ref. --Brainstem 1.89 (1.07–3.34) 0.028 ----1.59 (0.96–2.64)0.073Cerebellar 3.49 (1.9–6.43) < 0.001 ----2.56 (1.46–4.5) 0.001 Motor1.14 (0.68–1.92)0.610------Optic1.31 (0.82–2.10)0.252------ Migraine 1.49 (0.93–2.40)0.099------ Addiction 2.21 (1.37–3.57) 0.001 ---- - - Smoking 2.31 (1.46–3.64) < 0.001 3.58 (2.1–6.1) < 0.001 3.44 (2.03–5.84) < 0.001 2.21 (1.39–3.51) 0.001 Drugs (interferons) Ref. Ref. Ref. Ref. Ref. Ref. Ref. Ref. Dimethyl Fumarate0.98 (0.44–2.18)0.9520.91 (0.4–2.06)0.8161 (0.44–2.27)0.993--Fingolimod1.03 (0.44–2.45)0.9380.82 (0.34–1.97)0.6590.95 (0.4–2.27)0.908--Glatiramer Acetate0.69 (0.21–2.23)0.5300.8 (0.24–2.6)0.7060.78 (0.24–2.55)0.685--None--------Teriflunomide1.44 (0.74–2.80)0.2861.44 (0.73–2.84)0.2881.52 (0.77–2.99)0.225--Monoclonal antibodies 1.67 (1.13–2.47) 0.010 0.71 (0.4–1.25)0.2391.02 (0.56–1.87)0.938-- Association between study variables and BPD in MS subgroup Drugs (interferons) Discussion To the best of our knowledge, this is the first large-scale study to evaluate and compare the prevalence of BPD in NMOSD. The existing literature has highlighted an increased risk of psychiatric disorders—particularly BPD—among MS patients compared to the general population. Our findings are consistent with previous reports on the overall prevalence of BPD in MS and further demonstrate comparable rates among NMOSD patients. Moreover, our findings suggest that in clinically realistic scenarios—where EDSS varies naturally by subtype— SPMS patients had higher odds of BPD than patients with other MS subtypes. Because of the cross-sectional design, this finding should be interpreted as an association rather than evidence of a causal relationship. The observed prevalence of BPD in both MS and NMOSD aligns with earlier systematic reviews, where the pooled global prevalence of BPD in MS ranged between 2.95 and 5.83 per 100 patients [2, 3, 18]. Similarly, the higher risk of BPD in women was not immediately apparent prior to adjustment for other predictors, as observed in previous studies [2], but the adjusted estimate was in line with previous reports on matched Canadian population-based studies [4], suggesting a potentially higher prevalence of BPD among women with MS, in contrast to the relatively equal sex distribution seen in the general population. Furthermore, our estimated BPD prevalence (307 per 10,000) supports previous reports of elevated BPD risk in MS and NMOSD populations, particularly when compared to recent national prevalence estimates for BPD in Iran (78 per 10,000, 95% UI: [66.7–91.5]) [19]. Interestingly, we did not find an association between the presence of active brain lesions in conventional MRI and BPD; however, cerebellar symptoms at disease onset were more common among BPD patients. This aligns with prior evidence suggesting that early microstructural changes in the cerebellum, detectable via diffusion tensor imaging (DTI) rather than visible lesions, are associated with both longer disease duration and higher EDSS scores in MS patients [20]. The cerebellum is not only involved in motor coordination but also plays a critical role in higher-order cognitive and affective processing. Functional connectivity studies have shown that cerebellar pathways are intricately linked with prefrontal, temporal, parietal, and limbic regions—networks involved in executive functioning, emotion regulation, attention, language, and working memory [21, 22]. While cerebellar architecture is invariant throughout its cortex, type 2 BPD is characterized by regional gray matter reduction in Lobule IX and Crus-II of cerebellum, indicating potential role of the cerebellum in regulating social cognition [23]. Posterior lobe lesions in cerebellar cognitive affective syndrome and dysfunction in cerebello-thalamo-cortical loops have also been implicated in various neuropsychiatric and affective syndromes, including depression, anxiety, and mood dysregulation, which overlap phenomenologically with BPD features [24]. Be that as it may, we demonstrated that the higher risk of BPD associated with cerebellar symptoms to be independent of both disease duration and disability, suggesting that other unmeasured factors may be involved. The spatial distribution of brain and spinal cord lesions in relation to psychiatric symptoms in CNS ADD and their relationship with disease pathophysiology are therefore among topics of interest which should be further investigated in future studies. Despite strong evidence of a higher prevalence of psychiatric disorders in both MS and NMOSD [25, 26], the underlying pathophysiological mechanisms behind these associations remain poorly understood. Affective disorders are among the most frequently reported psychiatric conditions in these populations. The pooled prevalence of depression in particular was noted to be as high as 23.7% [3]. NMOSD patients may be prone to further psychiatric disorders as depression was reported to be as high as 30–50% in NMOSD patients and these patients were twice as likely to suffer from Major depressive disorder (MDD) compared to pwMS [13]. This is a critical concern, as depression as a core component of BPD may complicate diagnosis and treatment and is associated with poor adherence to DMTs [27]. BPD has been associated with elevated levels of pro-inflammatory cytokines, such as interleukin-6 (IL-6) and tumor necrosis factor (TNF), as well as an increased CD4+/CD8 + T-cell ratio—paralleling the immune activation seen in MS. DTI microstructural changes in corpus callosum are also seen with both MS & BPD, although MS patients typically show more pronounced abnormalities in the posterior corpus callosum, whereas individuals with BPD often exhibit significant changes in the anterior corpus callosum, particularly the genu, which is also involved in emotional regulation [28, 29]. Although, the literature has suggested shared genetic components in HLA immunophenotype of MS & BPD, no overlapping haplotypes were reported [30, 31]. On the other hand, early clinical evidence in NMOSD suggests a bidirectional relationship with co-occurring mood disorders, whereby immunosuppressive treatment is associated with concurrent improvement in affective symptoms and neurological function [32]. Taken together, these findings are consistent with a possible multifactorial relationship involving neuroinflammation, immune dysregulation, and white matter abnormalities in BPD among MS and NMOSD populations. Understanding the overlapping biological pathways may not only clarify disease mechanisms but also improve early identification and tailored management of psychiatric comorbidities in neuroinflammatory diseases. Although the present study was not designed to evaluate the longitudinal course or treatment response of BPD, MS and NMOSD may still influence its clinical expression and management. Overlapping symptoms such as fatigue, sleep disturbance, cognitive slowing, and affective lability may complicate early recognition of bipolar episodes in patients with CNS demyelinating disease. In addition, relapse treatment with corticosteroids may destabilize mood in vulnerable individuals, whereas disability burden and recurrent inflammatory activity may worsen psychosocial functioning and adherence to treatment. Standard bipolar treatments, including mood stabilizers and atypical antipsychotics when indicated, should be used with attention to neurological comorbidity, drug interactions, and steroid-related mood effects. For these reasons, a multidisciplinary approach is warranted, including routine psychiatric screening, careful monitoring during corticosteroid exposure, and timely collaboration between neurology and psychiatry when mood symptoms emerge. However, whether MS or NMOSD directly alters the long-term prognosis of BPD remains uncertain and should be addressed in longitudinal studies. Our study has several limitations. We were unable to further breakdown BPD cases into subtypes I and II and did not stratify NMOSD analyses further given the limited number of BPD cases in this group. Additionally, psychiatric diagnoses were based on DSM-IV criteria using SCID-I, as DSM-V-based structured interviews were not routinely implemented during the study period. However, diagnostic criteria for BPD have remained largely consistent between DSM-IV and DSM-V, and this is unlikely to have materially influenced prevalence estimates. EDSS scoring, though widely used, may also vary by examiner. Moreover, our NMOSD cohort was restricted to AQP4-antibody–positive patients to ensure diagnostic homogeneity. As AQP4-negative NMOSD represents a clinically heterogeneous group, which may differ in underlying pathophysiology and clinical manifestations, our findings may not be generalizable to the full NMOSD spectrum. Finally, because this was a cross-sectional study, temporal ordering cannot be established, and the observed associations should not be interpreted as causal, especially in the absence of further adjustment for other confounding factors including certain drug categories affecting both CNS ADD and psychiatric symptomatology (e.g. corticosteroids). In particular, higher EDSS may reflect disease severity rather than act as a predictor of BPD. Therefore, the causal link between SPMS and BPD remains elusive and whether BPD accelerates the demyelinating course of RRMS patients or vice versa should be investigated in matched cohort studies integrating neuroimaging, longitudinal follow-up, and neurobiological profiling to better understand the mechanistic pathways linking neuroinflammation, structural changes, and psychiatric vulnerability in demyelinating diseases. Conclusion Our findings highlight a comparable prevalence of BPD in MS and NMOSD, but higher risk of SPMS with regards to BPD among MS disease courses. Our findings underscore the importance of substance abuse, smoking, current disability and cerebellar involvement as factors which help identify patients at higher risk for BPD. Psychiatric screening should be a routine part of clinical care in both disorders, particularly in SPMS patients. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements Not applicable. Abbreviations Multiple sclerosis Central nervous system Bipolar disorder Neuromyelitis Optica Spectrum Disorder Autoimmune demyelinating diseases Diagnostic and Statistical Manual of Mental Disorders Structured Clinical Interview for DSM-IV Axis I Disorders Expanded Disability Status Scale Disease-modifying therapies Patients with MS Relapsing-remitting MS Primary progressive MS Secondary progressive MS Diffusion tensor imaging Author contributions Omid Mirmosayyeb: Conceptualization, Validation, Writing – Review & Editing, Project administration. Saeed Vaheb: Data curation, Supervision, Supervision. Sepehr Aghajanian: Methodology, Validation, Formal analysis, Writing- Original Draft, Visualization. Fatemeh Mohammadifard: Writing – Original Draft. Vahid Shaygannejad: Data curation, Supervision. Funding The authors declare that no funds, grants, or other support were received during the preparation of this manuscript. The authors did not receive support from any organization for the submitted work. Data availability The dataset used to reach the conclusion for this work is available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate The study and its procedures were performed in accordance with the Declaration of Helsinki and its amendments and was approved by the Ethics Committee of Isfahan University of Medical Sciences (IR.ARI.MUI.REC.1401.061). All included patients provided written informed consent after a full explanation of study protocol. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement The dataset used to reach the conclusion for this work is available from the corresponding author on reasonable request.