Background: Metabolomic analysis aids in the identification of novel biomarkers by revealing the metabolic dysregulations underlying cardiovascular disease (CVD) aetiology. The aim of this study was to evaluate which metabolic biomarkers could add value for the prognosis of CVD events using meta-analysis.Methods: The PRISMA guideline was followed for the systematic review. For the meta-analysis, biomarkers were included if they were tested in multivariate prediction models for fatal CVD outcomes. We grouped the metabolites in biological classes for subgroup analysis. We evaluated the prediction performance of models which reported discrimination and/or reclassification statistics.Results: For the systematic review, there were 22 studies which met the inclusion/exclusion criteria. For the meta-analysis, there were 41 metabolites grouped into 8 classes from 19 studies (45,420 subjects, 5954 events). A total of 39 of the 41 metabolites were significant with a combined effect size of 1.14 (1.07-1.20). For the predictive performance assessment, there were 21 studies, 54,337 subjects, 6415 events. The average change in c-statistic after adding the biomarkers to a clinical model was 0.0417 (SE 0.008).Conclusions: This study provides evidence that metabolomic biomarkers, mainly lipid species, have the potential to provide additional prognostic value. Current data are promising, although approaches and results are heterogeneous.

中文翻译：

---

代谢组学生物标志物对心血管疾病风险的预测价值：系统评价和荟萃分析。

背景：代谢组学分析可揭示心血管疾病（CVD）病因的代谢异常，从而帮助鉴定新型生物标志物。这项研究的目的是通过荟萃分析来评估哪些代谢生物标记物可以为CVD事件的预后增加价值。方法：遵循PRISMA指南进行系统评价。对于荟萃分析，如果在多因素预测模型中对致命性CVD结果进行了测试，则包括生物标志物。我们将代谢物按生物学分类，进行亚组分析。我们评估了报告歧视和/或重新分类统计数据的模型的预测性能。结果：对于系统评价，有22项研究符合纳入/排除标准。对于荟萃分析，19个研究将45种代谢物分为41类（45,420个受试者，5954个事件）。41种代谢物中总共39种具有显着性，合并效应大小为1.14（1.07-1.20）。对于预测绩效评估，共有21项研究，54,337名受试者，6415个事件。将生物标志物添加到临床模型后，c统计量的平均变化为0.0417（SE 0.008）。结论：本研究提供了代谢组学生物标志物（主要是脂质种类）有可能提供附加预后价值的证据。尽管方法和结果是不同的，但当前数据是有希望的。337个主题，6415个事件。将生物标志物加入临床模型后，c统计量的平均变化为0.0417（SE 0.008）。结论：本研究提供了代谢组学生物标志物（主要是脂质种类）有可能提供附加预后价值的证据。尽管方法和结果是不同的，但当前数据是有希望的。337个主题，6415个事件。将生物标志物添加到临床模型后，c统计量的平均变化为0.0417（SE 0.008）。结论：本研究提供了代谢组学生物标志物（主要是脂质种类）有可能提供附加预后价值的证据。尽管方法和结果是不同的，但当前数据是有希望的。