We retrospectively enrolled patients diagnosed with MOGAD at the First Affiliated Hospital of Zhengzhou University from January 2019 to December 2024 in this study. MOGAD was diagnosed based on the 2023 International Consensus Diagnostic Criteria for MOGAD (13).

The inclusion criteria were as follows: (1) patients seropositive for MOG antibodies, as detected by a live cell-based assay (CBA) using HEK293 cells expressing full-length human MOG with a secondary antibody to human IgG1 and (2) patients with confirmed first-attack MOGAD.

The exclusion criteria were as follows: (1) coexistence of other diseases that may affect the Expanded Disability Status Scale (EDSS) score; (2) coexistence of diseases involving renal dysfunction, vitamin B12 dysfunction, hypothyroidism, or hemolysis; (3) treatment with an Hcy-lowering drug before admission; (4) use of drugs affecting Hcy levels, such as isoniazid, phenytoin, levodopa, medroxyprogesterone, and others, before admission (14); (5) use of corticosteroids or immunosuppressive therapies in the 6 months before admission; (6) incomplete data; and (7) missing follow-up data. The detailed selection process is shown in Figure 1.

This study was approved by the Ethics Committee of the First Affiliated Hospital of Zhengzhou University (2023-KY-0398) and was conducted according to the principles of the Declaration of Helsinki.

Clinical data, including age at disease onset, sex, comorbidities, treatments, laboratory test results, and cerebrospinal fluid (CSF) analysis results at admission, were collected. The EDSS scores at admission and follow-up were set as the initial EDSS scores and final EDSS scores, respectively, and were evaluated by an experienced neurologist according to the method previously reported (15). Follow-up data were obtained via annual clinic visits or telephone interviews every 3 months. The last follow-up date was March 31, 2025.

Blood samples were collected from patients after an overnight fasting period (from 7:00 am to 8:00 am the next day and before the administration of any treatments). Patients were tested for MOG antibodies using a cell-based assay. Hcy levels were detected by an automated chemistry analyzer (Roche Group, Sweden). Hyperhomocysteinemia (high hcy group) was defined as an Hcy concentration >15 μmol/L (16, 17). Serum folic acid and vitamin B12 levels were detected by chemiluminescent immunoassay on the Maglumi 2000 Plus analyzer (Shenzhen New Industries Biomedical Engineering Company Limited, China). Tests were conducted in accordance with the manufacturer’s protocols; the examiners were blinded to the patients’ diagnoses and clinical symptoms.

MRI scans were performed using a 3.0 T scanner (Philips Healthcare, Amsterdam, Netherlands). Sagittal T1-weighted images (T1WIs); axial T1WIs; T2-weighted images (T2WIs); axial/sagittal fast fluid-attenuated inversion recovery images; axial diffusion weighted images; apparent diffusion coefficient mapped images; contrast-enhanced axial, coronal, and sagittal T1WIs of the brain were analyzed. And sagittal T1WIs, sagittal T2WIs, axial T1WIs, and axial T2WIs of the spine were analyzed. Gadolinium glutamine was injected as the contrast agent.

The clinical outcomes of the study were disability—according to the final EDSS score—and relapse event—at the last follow-up. The final EDSS score was converted into categorical variables. To explore the association between Hcy levels and clinical outcomes, patients who exhibited different final EDSS scores were divided into two groups. Patients with an EDSS score >3 were defined as having a poor recovery, while those with an EDSS score ≤3 were defined as having a good recovery. Relapse was defined as the presence of new-onset or recurrent neurological symptoms lasting for ≥24 h and resulting in an increase in the EDSS score of ≥0.5 points from the patient’s lowest score.

Normally distributed, continuous data are presented as mean±standard deviation. Continuous data with a non-normal distribution are presented as median (interquartile range). The Kolmogorov–Smirnov test was used to test the normality of the continuous data. Categorical variables are expressed as frequencies (percentages, %). The differences between groups were analyzed using the Student’s t-test and Wilcoxon test, for normally and non-normally distributed data, respectively. Categorical data were compared using the chi-square test when comparing numbers ≥5 or Fisher’s exact test when comparing numbers <5. Univariate logistic regression analysis was used to screen factors that might affect recovery in first-attack MOGAD. Variables with a significance level of p < 0.1 in the univariate logistic regression analysis were included in the basic multivariate logistic regression analysis model. Variables with p < 0.1 in the univariate logistic regression analysis or variables that may have had an impact on the final EDSS scores (including age at onset, sex, and initial EDSS scores) and factors that could affect Hcy levels (including levels of folic acid and vitamin B12) were included in the adjusted multivariate logistic regression analysis model to analyze the independent effects of Hcy levels on recovery. The correlation between Hcy levels and final EDSS scores was analyzed using Spearman’s correlation analysis. Kaplan–Meier analysis was performed to analyze the effect of Hcy levels on the timing of MOGAD relapse. A univariate Cox proportional hazards model was used to screen variables that might affect relapse in first-attack MOGAD. Variables with a significance level of p < 0.1 in the univariate Cox regression analysis were included in the basic multivariate Cox regression analysis model. Variables with p < 0.1 in the univariate Cox regression analysis or variables that may have had an impact on relapse (including age at onset, sex, and initial EDSS scores) and factors that could affect Hcy levels (including levels of folic acid and vitamin B12) were included in the adjusted multivariate Cox regression analysis model to analyze the independent effects of Hcy levels on relapse.

All statistical analyses were performed using SPSS (version 26.0; IBM, Armonk, NY, USA) and the diagram was generated with GraphPad Prism 8. Statistical significance was set at p < 0.05.

After inspecting the database of the First Affiliated Hospital of Zhengzhou University between January 2019 and December 2024, 101 patients were diagnosed as having MOGAD, with 70 patients meeting the inclusion criteria. Patients were divided into two groups according to serum Hcy levels: the high Hcy group (serum Hcy levels >15 μmol/L; n = 33) and the normal Hcy group (serum Hcy levels ≤15 μmol/L; n = 37). We compared the demographic and clinical characteristics of the two groups. As shown in Table 1, we found no differences in age, smoking status, alcohol use, and the prevalence of comorbidities—including hypertension, diabetes, coronary heart disease, anxiety/depression, malignancy, trauma, and autoimmune diseases—between the two groups (p > 0.05). The proportion of males and serum Hcy levels were higher, and folic acid levels were lower, in the high Hcy group than in the normal Hcy group (p < 0.05).

As shown in Table 2, no differences were found between the groups in intracranial pressure, proportion of brain lesions, optic nerve involvement, and spinal cord lesion segments, nor were differences found between the groups in levels of leukocytes, erythrocytes, hemoglobin, alanine aminotransferase, creatine, total cholesterol, triglycerides, high-density lipoprotein, low-density lipoprotein, cerebrospinal fluid leukocytes, or cerebrospinal fluid proteins (p > 0.05).

Initial EDSS scores were higher in the high Hcy group than in the normal Hcy group (p = 0.038). Patients received different treatments, including corticosteroids, immunoglobulin, and immunosuppressants (azathioprine, mycophenolate mofetil, and methotrexate), according to their clinical symptoms and financial situations. There were no significant differences in these parameters between the groups.

We found that the final EDSS scores were higher in the high Hcy group than in the normal Hcy group, indicating higher levels of disability in the high Hcy group (p = 0.015) (Figure 2A). Univariate analysis indicated that serum Hcy levels [odds ratio (OR) 1.124, 95% confidence interval (CI) 1.042–1.212, p = 0.002] and initial EDSS scores (OR 1.873, 95% CI 1.320–2.658, p = 0.000) were associated with poor clinical recovery (Table 3). In the basic multivariate logistic regression analysis model, Hcy levels (OR 1.104, 95% CI 1.010–1.206, p = 0.029) and initial EDSS scores (OR 1.825, 95% CI 1.240–2.686, p = 0.002) were related to poor clinical recovery of MOGAD. In the adjusted model, Hcy levels (OR 1.126; 95% CI 1.005–1.261, p = 0.04) and initial EDSS scores (OR 2.017, 95% CI 1.266–3.214, p = 0.003) remained associated with poor clinical recovery of MOGAD (Table 4). With Spearman’s correlation analysis, we found that Hcy levels (r = 0.03397; p = 0.004) were positively correlated with final EDSS scores (Figure 2B).

There were 15 cases relapsed in the study. Among these relapsed cases, 6 cases presented with myelitis, 2 cases presented with optic neuritis, 5 cases presented with brain lesions and 2 cases presented with encephalomyelitis.

The relapse rate was higher in the high Hcy group than in the normal Hcy group (p = 0.039) (Table 1). Kaplan–Meier survival analysis (Figure 3) showed that Hcy levels were a predictor of relapse in patients with MOGAD (log-rank test p = 0.029). Univariate analysis indicated that serum Hcy levels [hazard ratio (HR) 1.091, 95% CI 1.040–1.145, p = 0.000], age (HR 0.964, 95% CI 0.932–0.997, p = 0.034), and sex (HR 3.264, 95% CI 1.029–10.355, p = 0.045) were associated with the relapse of MOGAD (Table 5). In the basic multivariate Cox proportional hazards model, Hcy levels (HR 1.062, 95% CI 1.009–1.119, p = 0.022) were related to the relapse of MOGAD. In the adjusted model, Hcy levels (HR 1.088, 95% CI 1.020–1.161, p = 0.022) remained related to the relapse of MOGAD (Table 6), which indicated that Hcy levels may be a risk factor for predicting relapse.