Multiple sclerosis (MS) is an autoimmune disease predominantly mediated by T cells and marked by chronic inflammatory demyelination within the central nervous system (1). The pathophysiology of MS involves disruption of the blood–brain barrier, infiltration of immune cells, and production of myelin-specific autoantibodies, which collectively contribute to axonal damage and subsequent neurological dysfunction (2). Between 2013 and 2020, the global burden of MS increased, with the total number of affected individuals rising from 2.3 million to 2.8 million. In 2020, the estimated global prevalence was 36 per 100,000 population (95% confidence interval [CI]: 35.87–35.95) (3). Notably, the incidence of MS exhibits significant geographic disparities, with European countries reporting the highest rates (4). Given the growing disease burden of MS, conducting etiological research and implementing targeted preventive measures are essential.

Although the precise etiology of MS remains unclear, a growing body of evidence highlights the significant involvement of environmental exposures in MS development (5). Among modifiable risk factors, dietary carbohydrate intake has garnered increasing attention due to its potential role in modulating inflammation, a key mechanism in the onset of MS (6, 7). For instance, refined sugar intake is positively correlated with inflammatory biomarkers, specifically hypersensitive C-reactive protein (hs-CRP) and Interleukin-6 (IL-6), both of which are key players in systemic inflammation (8). Conversely, consumption of dietary fiber and whole grains has been linked to reduced inflammation, a benefit partly attributed to the production of short-chain fatty acids (SCFAs) through microbial fermentation in the gut (9, 10). Moreover, dietary carbohydrates may also have immune-modulating activities. For example, excessive fructose intake can exacerbate symptoms through gut microbiota dysbiosis and immune dysregulation in murine models of experimental autoimmune encephalomyelitis (EAE), which serves as an established animal model for MS (11). Similarly, high glucose levels promote the differentiation of T helper 17 (Th17) cells by activating the mTOR signaling pathway, thereby enhancing autoimmune responses in EAE (12). Collectively, these studies indicate that dietary carbohydrates can modulate both inflammatory and immune pathways in MS, with different types of carbohydrates exerting distinct biological effects.

To date, most nutritional epidemiological studies on MS have focused on overall dietary patterns (13–15), diet quality (16, 17), or specific food groups—such as fish (18), ultra-processed foods (19), fruits and vegetables (19), and processed red meat (20). Research on nutrient intake has largely emphasized fatty acids (21, 22) and vitamin D (23, 24). In contrast, limited population-level evidence exists regarding the role of carbohydrate intake in MS development. Notably, a recent prospective study using the UK Biobank (UKB) examined the association between the intake of grains, staples, vegetables, and fruits—major dietary sources of carbohydrates—and the risk of MS, but found no significant relationship (25). Given that epidemiological evidence on carbohydrates and MS remains scarce, we aimed to address this gap by leveraging the UKB dataset to comprehensively evaluate the associations between total carbohydrate intake, nine specific carbohydrate subtypes, and the risk of MS.

Of the 210,483 individuals free of MS at enrollment, 495 developed the disease over a median follow-up of 13.25 years (interquartile range [IQR]: 12.67–14.05). Among these incident cases, 358 (72.3%) were female. Table 1 presents the baseline demographic characteristics and macronutrient intakes across quartiles (Q1-Q4) of total carbohydrate intake. With increasing carbohydrate consumption, the proportions of male participants and individuals engaging in physical activity showed a gradual upward trend. Intakes of energy, fat, and protein also increased across the quartiles. Age was remarkably consistent across all four groups. The remaining characteristics did not follow a strict monotonic increasing or decreasing trend across quartiles but instead exhibited some variation.

The distribution of carbohydrate and its subtypes in both MS and non-MS populations is presented in Supplementary Table S2. Compared with the non-MS group, individuals with MS had significantly higher energy-adjusted intakes of carbohydrates (median: 260.65 vs. 256.01 g/d, p = 0.018), total sugars (median: 128.18 vs. 122.78 g/d, p < 0.001), fiber (median: 17.72 vs. 17.39 g/d, p = 0.012), fructose (median: 28.48 vs. 26.69 g/d, p < 0.001), glucose (median: 26.24 vs. 25.11 g/d, p = 0.004), and sucrose (median: 47.49 vs. 45.20 g/d, p = 0.001).

Associations of MS risk with each IQR increase in energy-adjusted carbohydrate intake are shown in Table 2. In the age- and sex-adjusted model (Model 1), per IQR increase in intake of energy-adjusted total carbohydrates (HR = 1.15, 95% CI: 1.03–1.28), total sugars (1.19, 1.08–1.31), fiber (1.14, 1.03–1.26), fructose (1.17, 1.06–1.29), glucose (1.15, 1.04–1.26), and sucrose (1.14, 1.04–1.25) were positively associated with MS risk, and all associations remained significant after false discovery rate (FDR) correction (all FDR-p < 0.05). After further adjustment for sociodemographic and lifestyle factors (Model 2), we observed that the associations for these carbohydrate types with MS risk were either slightly strengthened or remained similar, and all remained statistically significant (all FDR-p < 0.05). In the fully adjusted model, which included energy-adjusted fat and protein intake (Model 3), total carbohydrates (HR = 1.21, 95% CI: 1.05–1.40), total sugars (1.23, 1.10–1.38), fiber (1.20, 1.08–1.33), fructose (1.25, 1.12–1.39), and glucose (1.20, 1.08–1.34) continued to show positive associations with MS risk (all FDR-p < 0.05), with no substantial changes in effect sizes compared to Model 2. However, energy-adjusted sucrose was no longer statistically significant after FDR correction (FDR-p = 0.052).

In addition to analyzing carbohydrate intake as a continuous variable, we examined the associations between the five aforementioned carbohydrate types and MS by categorizing their intake into quartiles (Q1–Q4). As shown in Supplementary Table S3, participants in the third quartile (Q3) of energy-adjusted total carbohydrate intake exhibited a significantly higher MS risk compared to those in the lowest quartile (Q1) (HR = 1.41, 95% CI: 1.08–1.84) in the fully adjusted model. Energy-adjusted total sugars also demonstrated a positive association (Q4 vs. Q1: HR = 1.71, 95% CI: 1.29–2.27). Higher fiber intake corresponded to increased MS risk (Q4 vs. Q1: HR = 1.43, 95% CI: 1.10–1.86), and similar positive associations emerged for fructose (Q4 vs. Q1: HR = 1.62, 95% CI: 1.22–2.15) and glucose (Q4 vs. Q1: HR = 1.40, 95% CI: 1.06–1.86). These quartile-based findings align with the per IQR increase analysis, reinforcing a positive link between carbohydrate intake and MS risk.

We conducted several sensitivity analyses and found that the majority of the results remained stable, with minimal changes in effect sizes and sustained statistical significance. Notably, when the analysis was restricted to participants who completed at least three 24-h dietary assessments, the association between dietary fiber and MS was of borderline statistical significance, while the HR remained largely unchanged (HR = 1.19, 95% CI: 1.00–1.41; p = 0.056). For other carbohydrate types (total carbohydrates, total sugars, fructose, and glucose), greater variability in effect estimates was observed in this sensitivity analysis, although statistical significance was largely maintained. This increased variability may be attributed to the substantial reduction in sample size resulting from the inclusion criterion of at least three dietary assessments (Supplementary Table S4).

Based on these findings, we plotted RCS curves to visualize the dose–response relationships between carbohydrate intake and MS risk, as shown in Figure 1. Total carbohydrates, total sugars, fiber, fructose, and glucose intake all showed positive associations with MS risk (all P _overall < 0.05). Furthermore, each relationship was found to be linear (all P _nonlinear > 0.05).

The stratified analysis revealed that the associations of total carbohydrates, total sugars, fiber, fructose, and glucose with MS remained statistically significant only in females, but not in males. In age-stratified analyses, significant associations were observed for total carbohydrates and total sugars only among participants younger than 60 years old, whereas no significant associations were found in those aged 60 or older. In contrast, the strength of the associations for fiber, fructose, and glucose with MS appeared to be stronger in the older age group (≥60 years) (Figure 2).

Leveraging data from the UKB, this prospective cohort study provides the comprehensive evidence linking dietary carbohydrate intake and its subclasses with incident MS risk. Higher intakes of total carbohydrates, total sugars, fiber, fructose, and glucose were significantly associated with increased MS risk, with dose–response analyses revealing a linear association between these carbohydrate intake and MS incidence. Furthermore, these associations exhibited variations by sex and age.

Our findings indicate that a higher intake of carbohydrates is associated with an increased risk of MS, consistent with earlier reports. For example, the participants with MS were found to consume more carbohydrates than healthy individuals in a previous case–control study (50). Similarly, following a Mediterranean diet—which is naturally low in refined carbohydrates—has been linked to a lower risk of MS (51). Intervention studies further support the importance of carbohydrate quality: substituting refined grains with whole grains such as rye and whole-wheat pasta was shown to reduce levels of IL-6, an inflammatory molecule implicated in MS (52). Collectively, these findings suggest that both the quantity and quality of carbohydrates may influence MS risk.

In addition to total carbohydrate intake, we found that high sugar intake was significantly associated with increased MS risk. This aligns with mechanistic insights from a animal study, which indicated that excessive sugar may disrupt gut bacteria, particularly by depleting butyrate-producing bacteria that are crucial for maintaining both intestinal and blood–brain barrier integrity (53). Such dysbiosis could contribute to neuroinflammation and immune dysregulation, potentially explaining our observations. In particular, fructose and glucose—simple sugars abundant in sweetened foods and beverages—were strongly associated with MS risk. Experimental studies have shown that chronic high-fructose intake can induce hippocampal neuroinflammation, gliosis, and neuronal loss in mice (54), while glucose has been shown to promote Th17 cell differentiation, a process known to exacerbate neuroinflammatory responses (12). Nevertheless, further human studies are needed to clarify the specific roles of different sugars in MS pathogenesis.

From a public health perspective, our study highlights the potential of dietary modification as a population-wide strategy for reducing the risk of MS. The observed association between the intake of carbohydrates and sugars—particularly fructose and glucose—and MS incidence suggests that nutritional interventions could contribute meaningfully to the primary prevention of MS. These findings are consistent with new WHO guidelines strongly recommending that adults and children limit their daily intake of free sugars (defined as monosaccharides and disaccharides added to foods and beverages) to less than 10% of total energy intake to lower the risk of overweight, obesity, and dental caries (55). By extending the relevance of sugar reduction to neurological health, our results support the integration of dietary recommendations into public health policies aimed at preventing chronic immune-mediated diseases.

Interestingly, although fiber is generally considered beneficial due to its anti-inflammatory and gut-health promoting properties, our results showed a positive association with MS risk. This unexpected finding may be due to the different types of fiber. Dietary fibers are commonly classified into soluble and insoluble types based on their water solubility. Soluble fibers—such as inulin and guar gum—can be fermented by colonic microbiota, yielding metabolites including short-chain fatty acids like butyrate (56). In contrast, insoluble fibers, such as cellulose and hemicellulose, primarily enhance intestinal motility and are less readily utilized by gut microbes (57). Notably, emerging evidence suggests that different types of fiber may have divergent effects on health. For instance, previous study has reported that high doses of soluble fibers may promote colorectal carcinogenesis in mice, while insoluble fibers did not show similar effects (58). More specifically, high intake of soluble fiber may induce gut microbiota dysbiosis—enriching potential pathogens and reducing beneficial commensals—thereby facilitating colorectal tumorigenesis. Additionally, soluble fiber can disrupt intestinal metabolism, leading to elevated fecal butyrate and serum bile acids, along with reduced fecal inosine. Although direct evidence linking soluble fiber to MS is limited, we hypothesized that the positive association observed in our study may reflect high intake of soluble fiber and subsequent microbial metabolite activity, which could influence MS pathogenesis through the gut–brain axis. This hypothesis warrants further experimental validation.

Stratified analysis by sex revealed that the significant association between carbohydrate intake and MS was observed only in females. From a statistical perspective, this may be attributed to the larger sample size in this subgroup compared to males, where the limited sample size might have resulted in insufficient statistical power to detect a significant HR. Furthermore, the observed differences in associations may also reflect underlying sex-specific physiological mechanisms. MS incidence is higher in women, particularly during reproductive age (20–49 years), with a female-to-male ratio reaching up to 3:1 (59, 60). Hormonal influences play a key role in immune modulation. Estrogen, exhibits a biphasic immunomodulatory effect: it stimulates immune responses at low concentrations but acts as a strong immunosuppressant during pregnancy (33). Moreover, women generally demonstrate stronger antibody and T cell-mediated immune responses, which further contribute to this sexual dimorphism (61). The inherently higher susceptibility of women to MS may explain why the observed positive association between carbohydrate intake and MS was more pronounced in females.

Our study has some limitations should be noted. First, the UKB cohort primarily comprises middle-aged and older White participants, which limits generalizability to early-onset MS and other populations. Second, residual confounding cannot be completely excluded, particularly with respect to factors such as gut microbiome composition, which warrants prioritization in future research. Third, despite the prospective design, lag-period analysis, and extensive covariate adjustment, reverse causality remains plausible. Individuals with early MS symptoms or a high-risk profile may have altered their dietary habits prior to enrollment. However, the robustness of our findings after excluding early cases and adjusting for baseline chronic conditions mitigates this concern. Fourth, dietary data were self-reported, introducing potential measurement error. Fifth, although the overall cohort was large, the number of incident MS cases was modest, limiting statistical power and highlighting the need for replication in larger cohorts. Sixth, the UKB does not distinguish fiber subtypes (e.g., soluble vs. insoluble), preventing clarification of whether the observed positive association with MS risk is driven primarily by soluble fiber. Finally, while this study identifies correlative associations, it does not provide experimental validation of the mechanisms by which dietary carbohydrates influence MS pathology, which is a crucial next step that we plan to undertake in subsequent research.

In conclusion, we found significant associations between higher intakes of total carbohydrates, total sugars, fiber, fructose, and glucose and an increased MS risk. These findings offer valuable insights for developing dietary interventions focused on carbohydrate regulation to mitigate MS risk, which may help reduce its public health burden.