Publication
Research Article
There should be systematic assessment for bipolar disorder and schizophrenia in people with multiple sclerosis.
Background: Schizophrenia (SCZ) and bipolar disorder (BD) have been investigated in people with multiple sclerosis, with findings varying in prevalence. Our aim was to determine the global pervasiveness of SCZ and BD in people with MS.
Methods: Our study followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 guidelines and was registered in the International Prospective Register of Systematic Reviews (PROSPERO). Searches were performed in PubMed, Google Scholar, Europe PubMed Central, and Web of Science from their inception through April 2025. R statistical software was used for sensitivity analysis, subgroup analysis, and meta-regression, which was employed to address heterogeneity.
Results: The pooled prevalence of SCZ in people with MS was 0.61% (95% CI, 0.25%-1.11%), and BD was 3.34% (95% CI, 2.04%-4.91%). Leave-one-out analyses showed minimal to mild variation in both estimates. The prevalence of BD was higher in people with early-stage MS or moderate disability and varied by sex ratio. For both SCZ and BD, significant geographic differences were identified, with higher prevalence reported in the Americas and Europe. The study period influenced both SCZ and BD prevalences, although study design affected BD estimates and diagnostic approach influenced SCZ. The meta-regression identified study period, sample size, sex ratio, and diagnostic method as sources of heterogeneity in BD, although SCZ heterogeneity was driven by diagnostic method and prevalence type.
Conclusions: SCZ and BD are clinically relevant comorbidities in people with MS, with significant heterogeneity across studies driven by various factors. BD appears to be more prevalent in people with early-stage MS and in those with moderate disability. These findings highlight the need for standardized assessments and further studies.
Multiple sclerosis (MS) is a chronic autoimmune disease of the central nervous system characterized by inflammation and demyelination.1 The global prevalence of MS has risen significantly over the past 3 decades.2 Recent estimates indicate that approximately 2.8 million people worldwide have MS, equating to a prevalence rate of 35.9 per 100,000 population.2 Although its primary effects are neurological, MS can also manifest with various neuropsychiatric symptoms, including psychoses.2 Indeed, people with MS may have an increased risk of developing schizophrenia (SCZ) or bipolar disorder (BD).3 BD is a severe mental illness characterized by significant mood fluctuations, including episodes of mania or hypomania alternating with depression.4 SCZ is a severe psychiatric disorder characterized by disruptions in thought processes, perceptions, emotional responsiveness, and behaviors, and is characterized by symptoms such as hallucinations, delusions, and cognitive impairment.5 Several hypotheses have been proposed, including immune dysfunction and inflammation, to explain the association between MS and SCZ/BD.3 Neuroimaging research has revealed that symptoms can arise from diverse brain networks, regardless of the underlying pathological substrate, and from disruptions resulting from damage to critical white matter pathways.3 The connection between MS and SCZ/BD appears to be intricate and multifaceted. Both SCZ and BD can significantly impair social and occupational functioning and reduce quality of life when they co-occur with MS. Untreated psychosis may also adversely affect adherence to MS treatment regimens, potentially leading to worse neurological outcomes.6 Recognizing and addressing SCZ and BD in people with MS is essential for
comprehensive care.
Given the increased risk of psychoses in the global population with MS, it is crucial to assess the prevalence of SCZ and BD. Although meta-analyses estimating the prevalence of BD are limited,4 no meta-analysis to date has yet focused on the prevalence of SCZ. The aim of this study is to estimate the prevalence of both SCZ and BD in people with MS, helping to fill this gap in the literature.
Protocol Research and Registration
We registered the study on ClinicalTrials.gov (NCT05888727) and the study protocol with PROSPERO (registration ID: CRD42024557493). This systematic review and meta-analysis followed the protocols outlined in the Cochrane Handbook for Systematic Reviews of Interventions and adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines.7 The clinical inquiry was formulated using the Condition, Context, and Population framework (CoCoPop)8 (Table 1).
Table 1. CoCoPop Framework
Search Strategy
We conducted an exhaustive literature search spanning PubMed (MEDLINE), Google Scholar, Europe PubMed Central, and Web of Science to locate pertinent studies from database inception through April 2025. Employing a combination of keywords and medical subject heading terms, we structured our search strategy using the Health Terminology/Ontology Portal for enhanced precision. Table 1 presents a comprehensive list of terms used in each research study regarding the relationship between SCZ/BD and MS. The search process was independently conducted by 2 researchers (I.K. and F.T.). Additional studies were identified by examining references from systematic reviews on SCZ/BD/psychoses in MS4,9,10 and through searches in gray literature databases and conference abstracts. Identified records were imported into Mendeley (Elsevier) to identify and remove duplicates. Eligibility criteria screening included independent assessment of both titles and abstracts, followed by full-text screening using Covidence (Future Evidence Foundation). In cases of discrepancies, a third investigator was consulted (E.E.) and discussions persisted until a consensus was reached.
Eligibility Criteria
We included all studies that met the following criteria: (1) adults diagnosed with MS, irrespective of subtype and (2) observational designs reporting diagnoses of SCZ and/or BD, including cohort studies with a defined look-back period for identifying preexisting SCZ/BD, case-control studies with representative MS samples, and cross-sectional studies. There were no limitations regarding the publication date or language.
We excluded studies that replicated results using the same population, as well as studies that selected participants based on psychiatric status to minimize selection bias. We also excluded studies specifically designed to assess the incidence or risk of psychiatric disorders in relation to MS and those that involved psychoses induced by corticosteroid treatments or other factors. Articles were excluded if they did not present original study data (eg, reviews, book chapters, case reports), focused on the prevalence of SCZ/BD in pediatric MS, or lacked sufficient data. We also excluded reports that included schizoaffective disorder.
Quality Assessment and Bias Risk
We assessed the quality of all included articles using the JBI critical appraisal checklist, which is the recommended tool for evaluating descriptive studies reporting prevalence data. This checklist uses a scoring system based on the number of yes responses obtained in alignment with its predefined set of questions.11 A JBI score above 70% indicates a low risk of bias; scores between 50% and 69% suggest a moderate risk; scores below 49% denote a high risk of bias.
Data Extraction
Data extraction was performed using Excel 2019 (Microsoft Corp) with a standardized data collection form. The required information was the author’s name, study publication date, date of search, study design, study location, number of people with MS, number of people with MS with SCZ/BD diagnoses, type of MS (if available), and type of prevalence (point, period, or lifetime). Point prevalence refers to cases identified at the time of assessment or index date. Period prevalence refers to cases occurring within a defined time interval. Lifetime prevalence refers to studies containing self-reported lifetime diagnoses or registry or administrative records indicating an ever-diagnosed status.
Data Analyses
Data analysis was performed using the meta package in R version 4.2.3 (R Foundation for Statistical Computing). To account for potential variation among studies stemming from differences in their populations and origins, we opted for a random-effects model with inverse-variance weighting rather than a fixed-effects model. Heterogeneity across studies was assessed using the Cochran Q test (with a significance level set at P < .1) and the I2 statistic. Heterogeneity levels were categorized as minimal (I2 value, 0%-25%), low (25%-50%), moderate (50%-75%), and high (> 75%). The significance for the pooled effect was set at P < .05. For statistical significance, when P = .05, the effect was deemed significant if the 95% CI did not include 0. The Freeman-Tukey double-arcsine transformation model was utilized to compute the combined prevalence rates and pooled incidence rates.12 We chose this method because the proportions varied significantly across groups. This transformation model calculates weighted pooled estimates and then applies back-transformation on these estimates to stabilize the within-study variance.12 Sensitivity analysis, subgroup analysis, and univariate meta-regression were performed when significant heterogeneity was observed. Subgroup analysis and meta-regression were conducted based on the following factors: study origin, study source (population based or hospital based), study period, mean age of patients at the moment of the study, sex, population size (< 500, 500-1000, 1001-2000, and > 2000 people with MS), study quality, disability, type of questionnaire, disease duration (DD), and sex ratio (female/male) (< 1, 1-2, and >2). The cutoff of 40 years was chosen based on the approximate median age distribution across the included studies and the typical age range of people with MS reported in the literature. The sex ratio was categorized into 3 groups: less than 1, indicating male-dominant MS samples; 1 to 2, indicating mild to moderate female predominance; and greater than 2, indicating strong female predominance. DD was categorized into less than 5 years, 5 to 10 years, and greater than 10 years to reflect clinically relevant stages of multiple sclerosis progression, including early inflammatory, intermediate, and chronic phases.
We assessed the risk of publication bias in analyses that included more than 10 studies by visually inspecting funnel plots and conducting an Egger test. The significance level for the Egger test was set at P < .05, indicating potential publication bias.13
Study Characteristics and Quality Assessment
Figure 1 illustrates our search strategy in accordance with PRISMA 2020. Our screening process identified 101 records for eligibility. Full-text screening ultimately led to a total of 26 studies, including 78,925 patients with MS.3,14-38 For BD, we excluded the study by Wang et al,39 as it used the same data source as Huang et al.3 We prioritized the inclusion of Huang et al3 due to its specific focus on MS. In contrast, Wang et al39 investigated systemic autoimmune diseases as a broad exposure group, with MS representing only a small subset of the study population. Similarly, we excluded Laroni et al40 due to overlap with Carta et al.18 Carta et al was included due to its structured lifetime Diagnostic and Statistical Manual of Mental Disorders (Fourth Edition)–based diagnoses, whereas Laroni et al relied on registry data without standardized psychiatric assessment and reported BD within broader comorbidity categories. We excluded Marrie et al 200930 and 201541 and retained only the 2021 study27 for BD. Marrie et al 201541 did not report the number of patients with SCZ/BD, which prevented estimation of prevalence. Therefore, we prioritized the 2021 population-based administrative cohort study for BD, as it provided validated diagnostic coding and reduced the risk of misclassification.27 However, SCZ-specific estimates were not available in this data set. Therefore, we used the Marrie et al 2009,30 which reported lifetime prevalence estimates. Johansson et al1 was also excluded, as it involved the same population as Castelo-Branco et al.26 Johansson et al1 is a bidirectional registry-based cohort study designed to estimate the relative risk of developing MS in patients with psychiatric disorders and vice versa. Overall, we incorporated 9 studies for SCZ3,14,22,24,26,28-30,37 and 23 for BD in our analysis.3,14-21,23-28,31-38
Figure 1. Search Strategy According to PRISMA 2020
Table S1 (all supplemental tables are available in a PDF at the end of the online article) summarizes the characteristics of the included studies. The majority of studies were cross-sectional and conducted from 1987 to 2024. The mean (SD) age ranged from 32 ± 11.5 to 57.5 ± 13.5 years. The JBI score indicated that 15 studies were of high quality with a low risk of bias, and 11 studies were of moderate quality with a moderate risk of bias. The most reported outcome was lifetime prevalence.
Prevalence of Bipolar Disorder
The overall pooled prevalence of BD in patients with MS was 3.34% (95% CI, 2.04%-4.91%) (I2 = 96%; P < .01) (Figure 2A). The funnel plot appeared asymmetric, suggesting potential publication bias (Figure 2B). However, the Egger test was not statistically significant (P = .41). Given the limited number of studies and low statistical power, publication bias cannot be excluded. The leave-one-out sensitivity analysis showed minimal variation in the pooled prevalence, ranging from 3.04% (95% CI, 1.86%-4.47%) after excluding Joffe et al36 to 3.56%(95% CI, 2.22%-5.17%) after excluding Edwards et al32 (Table S2). Excluding cohort, case-control, and cross-sectional studies yielded estimates of 3.08%, 3.55%, and 4.17%, respectively (Table S3). Heterogeneity remained high across all analyses.
Figure 2A. Forest Plot of Pooled Prevalence of Bipolar Disorder in People With MS
Figure 2B. Funnel Plot of Prevalence of Bipolar Disorder in People With Multiple Sclerosis
The subgroup analysis by geographic region showed the highest prevalence of BD in people with MS in North and South America and the lowest prevalence in Oceania and Asia, with a significant difference (P < .01) (Figure S1). No significant differences were observed between hospital-based and
population-based studies (P = .10), across age groups (P = .33), or between prevalence types (P = .28) or diagnostic tools for BD (P = .08), although heterogeneity remained high across all analyses (Figures S2-S5). Studies conducted before 2015 showed the highest prevalence of BD, whereas those from 2015 to 2020 showed the lowest (P = .03), with high heterogeneity persisting (Figure S6). Case-control designs yielded higher estimates with lower heterogeneity (P = .01) (Figure S7). The prevalence of BD was highest in studies with a sex ratio of 1:2, followed by those with a ratio greater than 2, and lowest in those with a ratio less than 1 (P < .01), with lower heterogeneity in the less-than-1 and 1:2 subgroups (Figure S8). The subgroup analysis based on DD revealed a significantly higher prevalence of BD among people with MS with a DD of less than 5 years compared with longer durations (P < .01) (Figure S9). Data from 6 studies showed a significantly higher prevalence of BD in people with moderate disability compared with those with low disability (P = .02) (Figure S10).
Prevalence of Schizophrenia
The overall pooled prevalence of SCZ in people with MS was 0.61% (95% CI, 0.25%-1.11%) with high heterogeneity (I2 = 94%; P < .01) (Figure 3). Leave-one-out sensitivity analysis showed mild variation in the pooled prevalence, ranging from 0.46% (95% CI, 0.25%-0.72%) after excluding Demakis et al14 to 0.70% (95% CI, 0.30%-1.25%) after excluding Marrie et al 200930 (Table S4). Heterogeneity remained consistently high across all analyses.
Figure 3. Forest Plot of Pooled Prevalence of Schizophrenia in People With MS
The subgroup analysis by region showed highest prevalence of SCZ in Europe and North America and the lowest in Asia and Oceania (P < .01), with persistent heterogeneity (Figure S11). No significant differences were observed between hospital-based and population-based studies (P = .44), across age groups (P = .50), between prevalence types (P = .18) or study designs (P = .21), although heterogeneity remained high across all analyses (Figures S12-S15). Studies conducted between 2020 and 2025 showed the highest prevalence of SCZ, whereas those conducted between 2015 and 2020 had the lowest prevalence (P < .01), with high heterogeneity (Figure S16). Studies that used medical record–based diagnostics to identify SCZ demonstrated a higher prevalence compared with those relying on structured clinical interviews (P < .01) with low heterogeneity for medical interviews (Figure S17). However, subgroup analyses were not conducted according to sex ratio, DD, or disability due to missing data.
Univariate Meta-Regression Analysis
The meta-regression analysis showed that study period, sample size, sex ratio, and diagnostic method significantly influenced prevalence estimates for BD (Table S4). The observed heterogeneity in the prevalence of SCZ may be attributable to variations in diagnostic methods and the type of prevalence reported (Table S5).
Although prior meta-analyses have examined BD in MS, to our knowledge, this is the first to estimate the prevalence of SCZ and provides updated pooled estimates for BD. Our analysis yielded a pooled prevalence of 0.61% (95% CI, 0.25%-1.11%) for SCZ and 3.34% (95% CI, 2.04%-4.91%) for BD. The Global Burden of Disease 2021 estimates indicate that the prevalence of BD and SCZ in the general population is approximately 0.40% to 0.60%42 and 0.30%,43 respectively. Compared with our pooled results, these findings suggest a higher burden of these disorders among people with MS, underscoring the need to integrate mental health assessments into routine follow-up care. Several studies indicated that people with MS exhibited a risk of developing SCZ and BD.6,10,35 These findings suggest a possible association between MS and psychiatric comorbidities. However, as this is a prevalence meta-analysis, it does not allow for causal inference. Previous research has proposed several hypotheses that may explain this association.1,44-46 These include shared genetic susceptibility, with studies reporting specific human leukocyte antigen haplotypes associated with co-occurring MS and BD,44,47 and genome-wide loci potentially shared between MS and SCZ.45In addition, neuroinflammatory processes, such as microglial activation, cytokine dysregulation, and blood-brain barrier dysfunction, have been implicated in both mood disorders and MS.45,46 Structural brain changes, including white matter abnormalities and cortical gray matter alterations observed in MS and BD, have also been suggested as overlapping neurobiological features.48,49 Importantly, these mechanisms are based on prior studies and not the present analysis. These hypotheses may guide future longitudinal and mechanistic research on MS and psychiatric disorders.
Our results are broadly consistent with prior evidence. The most recent meta-analysis on BD, published in 2026, reported a prevalence of 3.20% (95% CI, 1.80%-4.60%) with high heterogeneity (I² = 97%),50 whereas a meta-analysis in 2021 estimated a prevalence of 2.95% (95% CI, 2.12%-4.09%) with similarly high heterogeneity (I² = 95%).4 The differences in prevalence estimates are likely attributable to variations in inclusion criteria and diagnostic tools. The 2026 meta-analysis50 included Meier et al,51 which reported incidences of SCZ and BD during MS follow-up rather than prevalence, potentially introducing selection bias when pooled with prevalence studies. Overlapping populations were insufficiently addressed, particularly among McKay et al31 and Berrigan et al.52 Possible overlap was also noted between Brenner et al53 and Castelo-Branco et al,26 as well as between Uguz et al15 and Uca et al.54 In our analysis, Fisk et al22 was retained for SCZ, and McKay et al31 for BD, as the prevalence of SCZ was not reported. The study by Berrigan et al52 was excluded due to its focus on quality-of-life outcomes rather than prevalence. The Brenner et al53 study was replaced by a more recent study from the same Swedish cohort (Castelo-Branco et al),26 which reported both BD and SCZ outcomes. Studies from the same Turkish center (Uguz et al15 and Uca et al54) were also evaluated, noting that Uca et al reported no BD cases in people with MS. The 2026 meta-analysis50 included only 6 people with BD from Ron et al,38 excluding people identified during the interview; however, we addressed this issue by including all reported people and performing subgroup analyses by prevalence type. The second meta-analysis (2021)4 incorporated both the 2009 and 2015 Marrie et al30,41 studies, which relied on the same data source, potentially introducing publication bias. In contrast, we retained only the study published in 2009 for SCZ.27
Overall, compared with prior meta-analyses, our meta-analysis provides a broader psychiatric epidemiological perspective by encompassing both SCZ and BD. On the other hand, the 20214 and 202650 meta-analyses, as well as the present meta-analysis, demonstrated substantial heterogeneity. Heterogeneity is widely supported in methodological literature, which emphasizes that high heterogeneity is common in prevalence syntheses and should be explored rather than eliminated.55 While random-effects models account for between-study variability, the high heterogeneity limits the generalizability and precision of the pooled estimates, which should therefore be interpreted cautiously as indicative trends rather than exact measures.
The leave-one-out analyses for both SCZ and BD demonstrated minimal to moderate variation in the pooled prevalence, suggesting that no single study disproportionately influenced the results. Nevertheless, some studies contributed to variability due to methodological differences. Demakis et al14 reported higher SCZ but lower BD prevalence, likely reflecting a nursing home–based population with MS and differing diagnostic ascertainment. Exclusion by study design yielded moderate variation in the pooled prevalence of BD. Removing cohort studies slightly lowered estimates (−0.26%), whereas excluding case-control (+0.21%) and cross-sectional (+0.83%) studies increased them. This pattern suggests that cross-sectional designs may underestimate BD, likely due to reliance on single time-point assessments and reduced sensitivity for episodic disorders. For SCZ, the higher prevalence observed in cohort studies is consistent with longitudinal follow-up, enabling more complete case ascertainment than cross-sectional approaches. Overall, the variations remained modest with overlapping 95% CIs, supporting the robustness of the findings. The subgroup analysis identified geographical variation and study period as significant sources of heterogeneity for both SCZ and BD. Geographical variation likely reflects differences in health care systems, diagnostic practices, and case ascertainment across regions for both SCZ and BD. Study period effects may reflect methodological changes rather than true epidemiological shifts, as earlier studies may have over- or underestimated prevalence due to less standardized assessments and selective populations. Notably, some geographical and temporal subgroups were based on a limited number of studies, which may have further affected the stability of the estimates. Consistent with our findings, Joseph et al reported a higher BD prevalence in people with MS in the Americas.4 Geographic variations in the prevalence of SCZ in MS continue to be debated.9 MS is more prevalent in northern and temperate regions, whereas SCZ is reported more frequently in Northern and Western Europe than in Southern Europe and Africa, highlighting substantial geographic heterogeneity.
The subgroup analysis by sex ratio (F:M) showed that the BD prevalence was highest in studies with a balanced sex distribution (F:M = 1:2). Joseph et al4 and Marrie et al56 reported higher BD rates in women, although this difference did not reach statistical significance. Sex ratio emerged as a source of heterogeneity in our meta-analysis. Indeed, female-predominant samples may yield higher BD estimates, as both BD and MS show greater prevalence in females. In contrast, male-skewed cohorts may underestimate BD burden due to the lower prevalence of both MS and BD in men. Balanced cohorts reduce selection bias and heterogeneity, yielding more stable estimates and helping explain between-study differences. For SCZ, most studies reported higher F:M ratios (> 2), except 1 with an F:M ratio of less than 1, suggesting a possible higher SCZ burden in women with MS, contrasting with the general population. A systematic review also reported higher SCZ incidence in women with MS.56 These findings highlight the need for further research to explore sex and gender differences in the prevalence and presentation of SCZ among people with MS.
DD emerged as a source of heterogeneity in the BD subgroup analysis, reflecting differences in MS clinical stages across studies. Results showed a higher BD prevalence in people with MS within the first 5 years, possibly related to early inflammatory activity, diagnostic uncertainty, and psychological stress, followed by a decline over time. However, it is worth noting that these estimates are based on only 2 studies of less than 5 years and 3 studies of more than 10 years, limiting their robustness.
Regarding disability, our findings suggest that people with MS with moderate disability may have a higher prevalence of BD compared with those with low disability. However, this result should be interpreted with caution. Notably, only 1 study specifically reported BD prevalence in people with MS with moderate disability, and most studies did not provide detailed disability stratification at the time of BD diagnosis. Therefore, residual confounding and reporting limitations may have influenced this observed association. Overall, these findings should be considered preliminary rather than definitive, and further studies are needed to confirm and strengthen these results.
Diagnostic methods showed no significant heterogeneity in the BD subgroup analysis, though meta-regression revealed slightly lower prevalence in medical records vs interviews. For SCZ, methods drove heterogeneity, with higher prevalence via medical records. The prevalence of BD appears relatively stable across diagnostic methods, whereas estimates of SCZ are more method dependent. For BD, structured or semistructured clinical interviews may enhance case detection by allowing systematic assessment of affective symptoms. Conversely, medical record–based ascertainment using International Classification of Diseases coding may miss undiagnosed or subthreshold BD cases, leading to underestimation of prevalence. For SCZ, diagnoses are typically more severe, chronic, and consistently documented in clinical records, which may explain the higher prevalence estimates observed in record-based studies. Therefore, differences in diagnostic tools appear to influence the pooled prevalence estimates, particularly for SCZ.
We found that sample size influenced BD prevalence in people with MS, with medium-sized studies reporting lower estimates than smaller studies. This may reflect a small-study effect driven by greater variability, selective sampling, or publication bias. Therefore, these findings should be considered exploratory rather than causal.
Prevalence type was not a significant source of heterogeneity for BD in subgroup analysis or meta-regression. For SCZ, meta-regression showed an association with prevalence type, with a significant effect for period prevalence. This may reflect its broader time window, increasing case detection compared with point prevalence. However, the effect was limited, suggesting a modest contribution to overall heterogeneity. Importantly, subgroup analyses showed no statistically significant differences between prevalence types for either SCZ or BD, indicating some inconsistency across analytical approaches.
Nevertheless, the results of the meta-regression for SCZ are based on a limited number of studies, reducing statistical power. Although univariable models were used to limit overfitting, the analyses remain exploratory and may not reliably identify true sources of heterogeneity. Thus, both null findings for most SCZ covariates and observed associations in BD should be considered hypothesis-generating rather than confirmatory.
Several factors were not significant sources of heterogeneity for SCZ or BD, including data source, study quality, and patient age. This suggests broadly comparable estimates across hospital- and population-based studies and limited influence of study quality or age on pooled prevalence. However, residual confounding cannot be excluded.
Clinical Implications
Our meta-analysis highlights the significant burden of SCZ and BD in individuals with MS, underscoring the need for routine screening for mood disorders, irrespective of prior psychiatric history. Given the observed variations in prevalence across geographic regions, study settings, and patient demographics, a tailored approach to psychiatric assessment is crucial. Early detection and management of SCZ and BD in people with MS can facilitate better disease control and improve overall quality of life. Integrating mental health evaluations into routine MS care through a multidisciplinary approach—considering regional, temporal, and demographic factors—can help optimize outcomes. Future research should further investigate the mechanisms underlying these associations to refine screening and treatment strategies.
Study Limitations
This study is the first to estimate the prevalence of SCZ in people with MS while providing an updated perspective on the prevalence of BD. One limitation is the inclusion of various study designs and the high heterogeneity observed in the prevalence findings, which may affect the credibility and generalizability of the pooled estimates. Other limitations also exist, particularly regarding the SCZ reports. Data on key MS characteristics, such as DD, Expanded Disability Status Scale scores, and MS subtype, were missing. In addition, the included studies had similar population characteristics, preventing meaningful subgroup analyses. Further, few of the included studies had control groups, preventing us from calculating ORs and estimating the risk of SCZ or BD in people with MS relative to the general population. The absence of data also made it impossible to compare prevalence rates between men and women with MS. Because most studies were hospital based and our meta-analysis showed high heterogeneity, the pooled estimates may underestimate the true prevalence of SCZ and BD in the broader population of people with MS. Therefore, caution is required in interpreting the results. Finally, the small number of included studies limits the power of the publication bias tests. Therefore, Egger test results and funnel plot interpretation should be interpreted with caution, and small-study effects cannot be excluded.
According to available studies, our meta-analysis suggests that SCZ and BD are prevalent in people with MS. Geographic variation was observed, with the highest prevalence for both disorders reported in the Americas. BD also appears more frequently in people with early-stage MS and those with moderate disability. However, these findings should be interpreted as preliminary due to methodological limitations. In the future, well-designed studies are needed to better characterize SCZ and BD in people with MS.
Conflicts of interest: The authors have disclosed no relevant conflicts of interest.
Johansson V, Lundholm C, Hillert J, et al. Multiple sclerosis and psychiatric disorders: comorbidity and sibling risk in a nationwide Swedish cohort. Mult Scler. 2014;20(14):1881-1891. doi:10.1177/1352458514540970
Khan G, Hashim MJ. Epidemiology of multiple sclerosis: global, regional, national and sub-national-level estimates and future projections. J Epidemiol Glob Health. 2025;15(1):21. doi:10.1007/s44197-025-00353-6
Huang YC, Chien WC, Chung CH, et al. Risk of psychiatric disorders in multiple sclerosis: a nationwide cohort study in an Asian population. Neuropsychiatr Dis Treat. 2021;17:587-604. doi:10.2147/NDT.S268360
Joseph B, Nandakumar AL, Ahmed AT, et al. Prevalence of bipolar disorder in multiple sclerosis: a systematic review and meta-analysis. Evid Based Ment Health. 2021;24(2):88-94. doi:10.1136/ebmental-2020-300207
Silverstein BH, Bressler SL, Diwadkar VA. Inferring the dysconnection syndrome in schizophrenia: interpretational considerations on methods for the network analyses of fMRI data. Front Psychiatry. 2016;7:132. doi:10.3389/fpsyt.2016.00132
Ellouz E, Ketata I, Abbes W, Gargouri H, Bader S. Psychoses in pediatric-onset multiple sclerosis: a case report and systematic review. J Child Adolesc Psychiatr Nurs. 2024;37(4):e70005. doi:10.1111/jcap.70005
Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71. doi:10.1136/bmj.n71
Hosseini M-S, Jahanshahlou F, Akbarzadeh MA, et al. Formulating research questions for evidence-based studies. J Med Surg Public Health. 2024;2:100046. doi:10.1016/j.glmedi.2023.100046
Arneth BM. Multiple sclerosis and schizophrenia. Int J Mol Sci. 2017;18(8):1760. doi:10.3390/ijms18081760
Ketata I, Ellouz E, Ketata A. Psychoses onset in multiple sclerosis: a systematic review and meta-analysis. Rev Neurol (Paris). 2025;181(6):483-502. doi:10.1016/j.neurol.2025.03.012
Zeng X, Zhang Y, Kwong JS, et al. The methodological quality assessment tools for preclinical and clinical studies, systematic review and meta-analysis, and clinical practice guideline: a systematic review. J Evid Based Med. 2015;8(1):2-10. doi:10.1111/jebm.12141
R Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing; 2013. Accessed June 15, 2025. https://www.R-project.org
Bowden J, Davey Smith G, Burgess S. Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression. Int J Epidemiol. 2015;44(2):512-525. doi:10.1093/ije/dyv080
Demakis GJ, Buchanan R, Dewald L. A longitudinal study of cognition in nursing home residents with multiple sclerosis. Disabil Rehabil. 2009;31(21):1734-1741. doi:10.1080/09638280902738664
Uguz F, Akpinar Z, Ozkan I, Tokgoz S. Mood and anxiety disorders in patients with multiple sclerosis. Int J Psychiatry Clin Pract. 2008;12(1):19-24. doi:10.1080/13651500701330825
Jun-O’Connell AH, Butala A, Morales IB, et al. The prevalence of bipolar disorders and association with quality of life in a cohort of patients with multiple sclerosis. J Neuropsychiatry Clin Neurosci. 2017;29(1):45-51. doi:10.1176/appi.neuropsych.15120403
Kotan VO, Kotan Z, Aydin B, et al. The relationship between psychopathology, psychosocial adjustment, social support and quality of life in multiple sclerosis. DergiPark Akad. 2019;5(1):20-28. doi:10.18621/eurj.382894
Carta MG, Moro MF, Lorefice L, et al. The risk of bipolar disorders in multiple sclerosis. J Affect Disord. 2014;155:255-260. doi:10.1016/j.jad.2013.11.008
Passarell MA, Otero-Romero S, Bufill E, et al. P415. Excess of neurological and psychiatric comorbidity in multiple sclerosis patients as compared to the general population in Catalonia, Spain. Poster presented at Joint ECTRIMS-ACTRIMS Meeting; October 25-28, 2017; Paris, France. Mult Scler. 2017;23(supp 3):169. doi:10.1177/1352458517731404
Kurnaz AHM, Ünlü MD, Demirci S. Prevalence of comorbidities in relapsing-remitting and secondary progressive multiple sclerosis. J Neurol Sci. 2019;405:199. doi:10.1016/j.jns.2019.10.1165
Weaver D, Charaf V, Reyes A, et al. P834. Psychological characterization of a cohort of multiple sclerosis patients in Chile. Poster presented at ECTRIMS; September 11-13, 2019; Stockholm, Sweden. Mult Scler. 2019;25(3):429-430. doi:10.1177/1352458519868080
Fisk JD, Morehouse SA, Brown MG, Skedgel C, Murray TJ. Hospital-based psychiatric service utilization and morbidity in multiple sclerosis. Can J Neurol Sci. 1998;25(3):230-235. doi:10.1017/s0317167100034065
Fatheddine M, Elfahiri F, Ennaciri Z, Adali I, Manoud F. Psychiatric comorbidities of multiple sclerosis: a cross-sectional study. SAS J Med. 2024;10(6):496-499. Accessed October 14, 2026. https://www.saspublishers.com/article/19991/
Silveira SL, Beydoun H, Tsai J. Association of multiple sclerosis with psychiatric disorders and homelessness among veterans in the United States Department of Veterans Affairs healthcare system. CNS Spectr. 2024;22:1-8. doi:10.1017/S1092852924002165
de Cerqueira AC, Semionato de Andrade P, Godoy Barreiros JM, Teixeira AL, Nardi AE. Psychiatric disorders in patients with multiple sclerosis. Compr Psychiatry. 2015;63:10-14. doi:10.1016/j.comppsych.2015.08.001
Castelo-Branco A, Chiesa F, Bengtsson CE, et al. Non-infectious comorbidity in patients with multiple sclerosis: a national cohort study in Sweden. Mult Scler J Exp Transl Clin. 2020;6(3):2055217320947761. doi:10.1177/2055217320947761
Marrie RA, Walld R, Bolton JM, et al. Effect of mood and anxiety disorders on health care utilization in multiple sclerosis. Mult Scler. 2021;27(9):1411-1420. doi:10.1177/1352458520963880
Lo LMP, Taylor BV, Winzenberg T, Palmer AJ, Blizzard L, van der Mei I. Change and onset-type differences in the prevalence of comorbidities in people with multiple sclerosis. J Neurol. 2021;268(2):602-612. doi:10.1007/s00415-020-10194-x
Rotstein D, Maxwell C, Tu K, et al. High prevalence of comorbidities at diagnosis in immigrants with multiple sclerosis. Mult Scler. 2021;27(12):1902-1913. doi:10.1177/13524585211031791
Marrie RA, Horwitz R, Cutter G, Tyry T, Campagnolo D, Vollmer T. The burden of mental comorbidity in multiple sclerosis: frequent, underdiagnosed, and undertreated. Mult Scler. 2009;15(3):385-392. doi:10.1177/1352458508099477
McKay KA, Tremlett H, Fisk JD, et al. Psychiatric comorbidity is associated with disability progression in multiple sclerosis. Neurology. 2018;90(15):e1316-e1323. doi:10.1212/WNL.0000000000005302
Edwards LJ, Constantinescu CS. A prospective study of conditions associated with multiple sclerosis in a cohort of 658 consecutive outpatients attending a multiple sclerosis clinic. Mult Scler. 2004;10(5):575-581. doi:10.1191/1352458504ms1087oa
Espinola-Nadurille M, Colin-Piana R, Ramirez-Bermudez J, et al. Mental disorders in Mexican patients with multiple sclerosis. J Neuropsychiatry Clin Neurosci. 2010;22(1):63-69. doi:10.1176/jnp.2010.22.1.63
Gerber B, Cowling T, Chen G, Yeung M, Duquette P, Haddad P. The impact of treatment adherence on clinical and economic outcomes in multiple sclerosis: real world evidence from Alberta, Canada. Mult Scler Relat Disord. 2017;18:218-224. doi:10.1016/j.msard.2017.10.001
Galeazzi GM, Ferrari S, Giaroli G, et al. Psychiatric disorders and depression in multiple sclerosis outpatients: impact of disability and interferon beta therapy. Neurol Sci. 2005;26(4):255-262. doi:10.1007/s10072-005-0468-8
Joffe RT, Lippert GP, Gray TA, Sawa G, Horvath Z. Mood disorder and multiple sclerosis. Arch Neurol. 1987;44(4):376-378. doi:10.1001/archneur.1987.00520160018007
Horton M, Rudick RA, Hara-Cleaver C, Marrie RA. Validation of a self-report comorbidity questionnaire for multiple sclerosis. Neuroepidemiology. 2010;35(2):83-90. doi:10.1159/000311013
Ron MA, Logsdail SJ. Psychiatric morbidity in multiple sclerosis: a clinical and MRI study. Psychol Med. 1989;19(4):887-895. doi:10.1017/s0033291700005602
Wang LY, Chiang JH, Chen SF, Shen YC. Systemic autoimmune diseases are associated with an increased risk of bipolar disorder: a nationwide population-based cohort study. J Affect Disord. 2018;227:31-37. doi:10.1016/j.jad.2017.10.027
Laroni A, Signori A, Maniscalco GT, et al. Assessing association of comorbidities with treatment choice and persistence in MS: a real-life multicenter study. Neurology. 2017;89(22):2222-2229. doi:10.1212/WNL.0000000000004686
Marrie RA, Fisk JD, Tremlett H, et al. Differences in the burden of psychiatric comorbidity in MS vs the general population. Neurology. 2015;85(22):1972-1979. doi:10.1212/WNL.0000000000002174
Jiang J, Chen Y, Lin Z, Liu S, Chen X, Chen Q. Global, regional, and national burden of bipolar disorder, 1990-2021: analysis of data from the Global Burden of Disease Study 2021. J Affect Disord. 2025;389:119638. doi:10.1016/j.jad.2025.119638
Feng Z, Shan X, Fan B, Yang L, Gong F. Global burden of schizophrenia in 204 countries and regions from 1990 to 2021 and machine learning-based projections to 2036: an analysis of the 2021 Global Burden of Disease Study. Front Psychiatry. 2026;16:1684168. doi:10.3389/fpsyt.2025.1684168
Kosmidis MH, Bozikas VP, Giannouli V, Karavatos A, Fokas K. Familial comorbidity of bipolar disorder and multiple sclerosis: genetic susceptibility, coexistence or causal relationship? Behav Neurol. 2012;25(4):341-349. doi:10.3233/BEN-2012-110198
Benros ME, Nielsen PR, Nordentoft M, Eaton WW, Dalton SO, Mortensen PB. Autoimmune diseases and severe infections as risk factors for schizophrenia: a 30-year population-based register study. Am J Psychiatry. 2011;168(12):1303-1310. doi:10.1176/appi.ajp.2011.11030516
Sun HL, Bai W, Li XH, et al. Schizophrenia and inflammation research: a bibliometric analysis. Front Immunol. 2022;13:907851. doi:10.3389/fimmu.2022.907851
Schiffer RB, Weitkamp LR, Wineman NM, Guttormsen S. Multiple sclerosis and affective disorder: family history, sex, and HLA-DR antigens. Arch Neurol. 1988;45(12):1345-1348. doi:10.1001/archneur.1988.00520360063013
Kutzelnigg A, Lucchinetti CF, Stadelmann C, et al. Cortical demyelination and diffuse white matter injury in multiple sclerosis. Brain. 2005;128(11):2705-2712. doi:10.1093/brain/awh641
Pape K, Tamouza R, Leboyer M, Zipp F. Immunoneuropsychiatry: novel perspectives on brain disorders. Nat Rev Neurol. 2019;15(6):317-328. doi:10.1038/s41582-019-0174-4
Mirmosayyeb O, Vaheb S, Shaygannejad A, Foroughi E, Yazdan Panah M, Shaygannejad V. Bipolar disorder in people with multiple sclerosis: an updated systematic review and meta-analysis. BMC Psychiatry. 2026;26(1):204. doi:10.1186/s12888-026-07839-y
Meier UC, Ramagopalan SV, Goldacre MJ, Goldacre R. Risk of schizophrenia and bipolar disorder in patients with multiple sclerosis: record-linkage studies. Front Psychiatry. 2020;11:662. doi:10.3389/fpsyt.2020.00662
Berrigan LI, Fisk JD, Patten SB, et al. Health-related quality of life in multiple sclerosis: direct and indirect effects of comorbidity. Neurology. 2016;86(15):1417-1424. doi:10.1212/WNL.0000000000002564
Brenner P, Alexanderson K, Björkenstam C, et al. Psychiatric diagnoses, medication and risk for disability pension in multiple sclerosis patients: a population-based register study. PLoS One. 2014;9(8):e104165. doi:10.1371/journal.pone.0104165
Uca AU, Uguz F, Kozak HH, Turgut K, Tekin G, Altas M, et al. Personality disorders in patients with multiple sclerosis: prevalence and association with depressive and anxiety disorders and clinical features. Neurol Asia. 2016;21(1):55-61.
Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ. 2003;327(7414):557-560. doi:10.1136/bmj.327.7414.557
Marrie RA, Reingold S, Cohen J, et al. The incidence and prevalence of psychiatric disorders in multiple sclerosis: a systematic review. Mult Scler. 2015;21(3):305-317. doi:10.1177/1352458514564487
