Scholarlink

Background <p>Febrile seizures (FS) represent the most common type of seizures in children; however, their exact pathogenesis remains incompletely understood. Currently, there is a lack of specific biomarkers for predicting FS occurrence, and existing prophylactic drug strategies remain controversial. Using untargeted metabolomics, this study investigates metabolic differences between children with FS and those with fever but without seizures (non-febrile seizures, NFS), aiming to elucidate the metabolic profile of FS and identify potential biomarkers, thereby providing new insights for clinical prediction and treatment.</p> Methods <p>Plasma samples were collected from 31 children with FS and 31 children with NFS. Untargeted metabolomic profiling was performed using high-performance liquid chromatography coupled with high-resolution mass spectrometry (HPLC-HRMS). Peak extraction and metabolite identification were conducted with the XCMS software. Differential metabolites were screened using both univariate and multivariate statistical analyses, followed by metabolic pathway enrichment analysis. A random forest algorithm was applied to construct a predictive model, and significantly altered metabolites were selected as candidate biological predictors.</p> Results <p>Children with FS exhibited significant metabolic disturbances across multiple pathways, including necroptosis, glycerophospholipid metabolism, linoleic acid metabolism, sphingolipid signaling, phagocytosis, ferroptosis, and sphingolipid metabolism. The random forest model identified 10 significantly altered metabolites as potential predictors: SM(d18:1/24:1), LysoPC(22:0/0:0), SM(d18:0/18:0), Cer(d18:1/16:0), LysoPC(24:0/0:0), 12,13-DHOME, diethanolamine, pantothenic acid, arachidonic acid, and 3-carbamoyl-2-phenylpropionaldehyde. The model demonstrated a predictive accuracy of 83% and achieved an area under the curve (AUC) of 0.98.</p> Conclusion <p>Patients with FS exhibited a distinct metabolic profile characterized by activated necroptosis, dysregulated lipid metabolism, and inflammatory imbalance. Metabolites such as arachidonic acid, lysophosphatidylcholines, and sphingolipids may serve as potential biomarkers and therapeutic targets. This study provides new metabolomic evidence for early prediction and targeted intervention of FS.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Metabolomic identification and analysis of potential biomarkers of febrile seizures

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年1月14日

查看全文
  • Haiting Tang,
  • Guilin Yuan,
  • Xiaowen Li,
  • Shaolun Pan,
  • Yaowen Liang,
  • Quan Yang,
  • Xiaoyan Gao

摘要

Background

Febrile seizures (FS) represent the most common type of seizures in children; however, their exact pathogenesis remains incompletely understood. Currently, there is a lack of specific biomarkers for predicting FS occurrence, and existing prophylactic drug strategies remain controversial. Using untargeted metabolomics, this study investigates metabolic differences between children with FS and those with fever but without seizures (non-febrile seizures, NFS), aiming to elucidate the metabolic profile of FS and identify potential biomarkers, thereby providing new insights for clinical prediction and treatment.

Methods

Plasma samples were collected from 31 children with FS and 31 children with NFS. Untargeted metabolomic profiling was performed using high-performance liquid chromatography coupled with high-resolution mass spectrometry (HPLC-HRMS). Peak extraction and metabolite identification were conducted with the XCMS software. Differential metabolites were screened using both univariate and multivariate statistical analyses, followed by metabolic pathway enrichment analysis. A random forest algorithm was applied to construct a predictive model, and significantly altered metabolites were selected as candidate biological predictors.

Results

Children with FS exhibited significant metabolic disturbances across multiple pathways, including necroptosis, glycerophospholipid metabolism, linoleic acid metabolism, sphingolipid signaling, phagocytosis, ferroptosis, and sphingolipid metabolism. The random forest model identified 10 significantly altered metabolites as potential predictors: SM(d18:1/24:1), LysoPC(22:0/0:0), SM(d18:0/18:0), Cer(d18:1/16:0), LysoPC(24:0/0:0), 12,13-DHOME, diethanolamine, pantothenic acid, arachidonic acid, and 3-carbamoyl-2-phenylpropionaldehyde. The model demonstrated a predictive accuracy of 83% and achieved an area under the curve (AUC) of 0.98.

Conclusion

Patients with FS exhibited a distinct metabolic profile characterized by activated necroptosis, dysregulated lipid metabolism, and inflammatory imbalance. Metabolites such as arachidonic acid, lysophosphatidylcholines, and sphingolipids may serve as potential biomarkers and therapeutic targets. This study provides new metabolomic evidence for early prediction and targeted intervention of FS.