Random-effects models are central to meta-analysis, yet the between-study variance is often underestimated when the number of studies is small. In such settings, confidence intervals become unduly narrow and fail to attain the nominal coverage probability. Although several small-sample corrections, including the Bartlett correction, have been developed under the normal-normal model, corresponding methodology for generalized linear mixed models (GLMMs) remains limited. This study proposes a unified framework for random-effects meta-analysis within the GLMM framework that relies exclusively on aggregate data and accommodates outcomes following any distribution in the exponential family, including the binomial, Poisson, and gamma distributions. To improve interval estimation with few studies, we develop a profile likelihood method with a simplified Bartlett correction (PLSBC), which refines the chi-squared approximation of the profile likelihood ratio statistic without requiring higher-order derivatives. We show theoretically that the proposed estimators preserve the consistency and asymptotic normality of the maximum likelihood estimators. Simulation studies demonstrate that the PLSBC yields nearly unbiased estimates and maintains nominal coverage across a variety of outcome types. Applications to three published meta-analyses with binomial, Poisson, and gamma outcomes indicate that the proposed approach provides robust and interpretable inference with few studies. The PLSBC therefore offers a practical and broadly applicable framework for random-effects meta-analysis when the number of studies is limited.

Random-effects meta-analysis is widely used for synthesizing results across studies, but the implicit assumption of normality in study-specific aggregate data is often violated. Such violations can lead to biased estimates and misleading conclusions, especially in meta-analyses with small studies or rare events. A prominent example occurs with the log-odds ratio, which exhibits bias that depends on the within-study sample sizes. We first show that conventional methods assuming normality fail to eliminate such biases, even as the number of studies increases. To overcome this limitation, we introduce a generalized linear mixed-effects model that uses only aggregate data, accommodating a wide range of outcome types, including binomial, Poisson, gamma, and other members of the exponential family, without requiring individual participant data. To enable valid interval estimation when the number of studies is small, we further propose a simple Bartlett correction for the test statistics. The proposed method yields consistent point estimators without relying on the normality assumption and achieves accurate interval coverage across diverse outcome types. It is particularly applicable in clinical and epidemiological research where only summary data are available, making it a practical alternative to conventional approaches. Simulation studies and applications to three published meta-analyses with binary, Poisson, and gamma outcomes demonstrate that the method provides reliable inference and maintains nominal coverage, thereby supporting sound decision-making and guideline development when only aggregate data are available.

Cholangiocarcinoma (CC) is a highly lethal malignancy that urgently requires reliable prognostic biomarkers. Although MUC1 expression and promoter methylation have been implicated in CC, the clinical significance of promoter methylation pattern composition, beyond average methylation levels, remains unclear. Here, we investigated the relationship between MUC1 promoter methylation heterogeneity, MUC1 mRNA expression, and prognosis in CC. We analyzed bisulfite amplicon sequencing data and mRNA expression of MUC1, DNA methylation-related enzymes (TET1, TET2, TET3, Dnmt1, and Dnmt3a), and tumor microenvironment stress markers in 131 CC tissues. In the neoplastic region, high MUC1 mRNA expression was associated with poor overall survival (HR = 0.131, 95% CI: 0.02 to 0.95, p = 0.042) and correlated with the abundance of completely unmethylated promoter patterns (r = 0.386, p < 0.001). Among the enzymes analyzed, only TET3 expression significantly correlated with the abundance of completely unmethylated patterns in the neoplastic region (Cohen’s f2 = 0.108, p = 0.009), suggesting a potential region-specific regulatory association. We visualized beta-diversity in methylation pattern composition using t-SNE and classified samples into two groups based on a linear decision boundary in the t-SNE space. This classification stratified prognosis independently of clinical factors (HR = 0.291, 95% CI: 0.06 to 0.94, p = 0.037; multivariate p = 0.021). These findings propose a novel, composition-based epigenetic stratification framework in CC, revealing that MUC1 promoter methylation pattern structure—rather than average methylation level—has prognostic relevance. Our results highlight the potential of pattern-resolved methylation profiling in the development of clinically applicable epigenetic biomarkers.

Meta-analyses are commonly used to provide solid evidence across numerous studies. Traditional moment methods, such as the DerSimonian-Laird method, remain popular in spite of the availability of more accurate alternatives. While moment estimators are simple and intuitive, they are known to underestimate the variance of the overall treatment effect, particularly when the number of studies is small. This underestimation can lead to excessively narrow confidence intervals that do not meet the nominal confidence level, potentially resulting in misleading conclusions. In this study, we improve traditional moment-based meta-analysis methods by incorporating Huber's M-estimation to more accurately capture the distributional characteristics of between-study variance. Our approach enables conservative parameter estimation, even when almost all existing methods lead to underestimation of between-study variance under a small number of studies. Additionally, by deriving the simultaneous distribution of overall treatment effect and between-study variance, we propose facilitating a visual exploration of the relationship between these two quantities. Our method provides more reliable estimators for the overall treatment effect and between-study variance, particularly in situations with few studies. Using simulations and real data analysis, we demonstrate that our approach always yields more conservative results compared to traditional moment methods, and ensures more accurate confidence intervals in meta-analyses.

Parkinson’s disease (PD) is a neurodegenerative movement disorder associated with a loss of dopamine neurons in the substantia nigra. The diagnosis of PD is sensitive since it shows clinical features that are common with other neurodegenerative diseases. In addition, most symptoms arise at the late stage of the disease, where most dopaminergic neurons are already damaged. Several studies reported that oxidative stress is a key modulator in the development of PD. This condition occurs due to excess reactive oxygen species (ROS) production in the cellular system and the incapability of antioxidants to neutralize it. In this study, we focused on the pathology of PD by measuring serum xanthine oxidase (XO) activity, which is an enzyme that generates ROS. Interestingly, the serum XO activity of patients with PD was markedly upregulated compared to patients with other neurological diseases (ONDs) as a control. Moreover, serum XO activity in patients with PD showed a significant correlation with the disease severity based on the Hoehn and Yahr (HY) stages. The investigation of antioxidant status also revealed that serum uric acid levels were significantly lower in the severe group (HY ≥ 3) than in the ONDs group. Together, these results suggest that XO activity may contribute to the development of PD and might potentially be a biomarker for determining disease severity in patients with PD.

ObjectivesThis study aimed to evaluate whether switching disease-modifying therapies (DMTs) from sphingosine-1 phosphate (S1P) receptor modulators to either natalizumab (NTZ) or dimethyl fumarate (DMF) could restore the effectiveness of SARS-CoV-2 mRNA vaccination in patients with multiple sclerosis (MS).MethodsThis study included 9 controls and 33 patients with MS: 7 patients treated with DMF, 7 patients treated with NTZ, 9 patients treated with S1P receptor modulators, and 10 patients who had switched DMTs from S1P receptor modulators to DMF or NTZ by the second vaccine dose. The patients who had switched DMTs were classified into two groups, based on whether their lymphocyte counts were above or below 1000/μL at the time of vaccination. In addition, relapses within 6 months after switching DMTs were also evaluated in these patients. Six months after the second dose of the vaccination, anti-SARS-CoV-2 spike antibodies were evaluated in all participants, and spike specific CD4+ T cells were also assessed in patients who had switched DMTs from S1P receptor modulators.ResultsPatients treated with S1P receptor modulators had lower levels of anti-SARS-CoV-2 spike antibodies than the controls and patients treated with DMF and NTZ. On the other hand, in patients who had switched DMTs from S1P receptor modulators, a recovery of lymphocyte counts above 1000/µL resulted in restored humoral and cellular immune responses to the vaccination. There were no neurological relapses in patients who had switched DMTs from S1P receptor modulators to NTZ.ConclusionSARS-CoV-2 mRNA vaccination is expected to be effective in patients whose lymphocyte counts have recovered due to switching DMTs from S1P receptor modulators. Switching DMTs from S1P receptor modulators to NTZ before vaccination may be beneficial in achieving efficacy for SARS-CoV-2 mRNA vaccination, with a reduced risk of relapse.

The utilization of real-world data is gaining momentum. However, when applying Cox regression model to real data containing time-dependent covariates, we are almost always forced to deal with incomplete data. We propose using Gaussian process regression technique to complete the missing values necessary to estimate the parameters of Cox regression model. Multiple imputation framework is also used to consider the uncertainty of imputed values. Utilizing the information of auxiliary variables are also discussed. Results of simulation study suggest that the combination of Gaussian process regression and multiple imputation method provides better estimates than standard method. The proposed method is simple and easy to implement, and the results are highly convincing.

Difference in the expression level of mucins between Cluster 1 and other clusters selected by cluster analysis of the methylation status of mucin genes MUC1, MUC2, and MUC4. Expression levels show relative quantification (log10).

LOOCV test and clinicopathological data

Repulsive guidance molecule A (RGMa) was originally identified as a neuronal growth cone–collapsing factor. Previous reports have demonstrated the multifunctional roles of RGMa mediated by neogenin1. However, the pathogenic involvement of RGMa in amyotrophic lateral sclerosis (ALS) remains unclear. Here, we demonstrated that RGMa concentration was elevated in the cerebrospinal fluid of both patients with ALS and transgenic mice overexpressing the mutant human superoxide dismutase1 (mSOD1 mice). Treatment with humanized anti-RGMa monoclonal antibody ameliorated the clinical symptoms in mSOD1 mice. Histochemical analysis revealed that the anti-RGMa antibody significantly decreased mutant SOD1 protein accumulation in the motor neurons of mSOD1 mice via inhibition of actin depolymerization. In vitro analysis revealed that the anti-RGMa antibody inhibited the cellular uptake of the mutant SOD1 protein, presumably by reinforcing the neuronal actin barrier. Collectively, these data suggest that RGMa leads to the collapse of the neuronal actin barrier and promotes aberrant protein deposition, resulting in exacerbation of the ALS pathology.

Prognosis prediction by machine learning classifier in the k-fold CV test. Cox proportional hazard regression analysis for the comparison between the positive and negative groups selected by each classifier. Red solid line: predicted high-risk group (positive), blue line: other groups (negative). (A) 3-fold cross-validation test. (B) 4-fold cross-validation test. The insets show p-value (p) and hazard ratio (HR) of each test.

Univariate and multivariate analyses for overall survival (Cox proportional hazard model).

Cox proportional hazard regression analysis on a comparison between Cluster 1 and other clusters as selected by cluster analysis of the methylation status of mucin genes MUC1, MUC2, and MUC4 in non-neoplastic regions (A) and neoplastic regions (B). Red solid line: Cluster 1, black dashed line: other clusters.

Cox proportional hazard regression analysis on a comparison between Cluster 1 and Cluster 2 selected by cluster analysis of mRNA expression levels of MUC1, MUC2, and MUC4 in non-neoplastic regions (A) and neoplastic regions (B). Red solid line: Cluster 1, black dashed line: Cluster 2.

Comparison of feature between training and test cohort.

Comparison of SVM predict ability among these kernels in LOOCV