Early Diagnosis of Bloodstream Infections Using Serum Metabolomic Analysis

Abstract

Background: Bloodstream infections (BSIs) pose a great challenge to treating patients, especially those with underlying diseases, such as immunodeficiency diseases. Early diagnosis helps to direct precise empirical antibiotic administration and proper clinical management. This study carried out a serum metabolomic analysis using blood specimens sampled from patients with a suspected infection whose routine culture results were later demonstrated to be positive.

Methods: A liquid chromatograph-mass spectrometry-based metabolomic analysis was carried out to profile the BSI serum samples. The serum metabolomics data could be used to successfully differentiate BSIs from non-BSIs.

Results: The major classes of the isolated pathogens (e.g., Gram-positive and Gram-negative bacteria) could be differentiated using our optimized statistical algorithms. In addition, by using different machine-learning algorithms, the isolated pathogens could also be classified at the species levels (e.g., Escherichia coli and Klebsiella pneumoniae) or according to their specific antibiotic-resistant phenotypes (e.g., extended-spectrum β-lactamase-producing and non-producing phenotypes) if needed.

Conclusions: This study provides an early diagnosis method that could be an alternative to the traditional time-consuming culture process to identify BSIs. Moreover, this metabolomics strategy was less affected by several risk factors (e.g., antibiotics administration) that could produce false culture results.

Keywords: bacteria; bloodstream infection; extended-spectrum β-lactamase; fungi; metabolomics.

Figure 1

Evaluation of the metabolomic method stability for the first fraction of the hydrophilic extract analysis: (a) PCA score plot of the QC and real samples; (b) coefficient of variation (CV%) distribution of metabolite intensities in QC samples; (c) time series plot of principal component 1 (PC1) of the QC samples after Pareto Scaling for PCA.

Figure 2

Exploring the BSI−affected metabolic pathways: (a) OPLS−DA score plot for BSI and nonBSI Samples; (b) metabolic pathway enrichment analysis highlighting the most perturbed pathways by infections; (c) bubble plot of metabolites with fold change > 2 and Q−value < 0.05 between BSI and non-BSI samples. VIP: variable importance.

Figure 3

Results of the four-class separation of the isolated strains: (a) the best-performing model with top 20 metabolites from 100 LASSO runs in pre-selection stage; (b) AUC and accuracy values with top 1-20 metabolites in the evaluation stage.

Figure 4

Separation of the Gram−positive and Gram-negative bacterial BSIs by XGBoost model: (a) t−SNE plot for classification of Gram−Positive and Gram−Negative bacteria; (b) confusion matrix of XGBoost model based on the four selected metabolite variables; (c) increase in AUC with the inclusion of more metabolites in the model.

Figure 5

Results of the identification of E. coli and K. pneumoniae BSIs: (a) comparison of AUC before and after optimization of XGBoost model (87 variables); (b) SHAP summary plot of TPE−optimized XGBoost model showing metabolite contributions for bacterial differentiation; (c) comparison of AUC before and after optimization of XGBoost model (7 variables).

Figure 6

Results of the model separation according to ESBL−producing phenotypes: (a) SHAP summary plot of XGBoost model based on the optimal parameter combination; (b) top 10 features identified by permutation feature importance (PFI); (c) comparison of AUC before and after optimization of XGBoost model (six variables).