Metabolic profiles in community-acquired pneumonia: developing assessment tools for disease severity

Abstract

Background: This study aimed to determine whether community-acquired pneumonia (CAP) had a metabolic profile and whether this profile can be used for disease severity assessment.

Methods: A total of 175 individuals including 119 CAP patients and 56 controls were enrolled and divided into two cohorts. Serum samples from a discovery cohort (n = 102, including 38 non-severe CAP, 30 severe CAP, and 34 age and sex-matched controls) were determined by untargeted ultra-high-performance liquid chromatography with tandem mass spectrometry (LC-MS/MS)-based metabolomics. Selected differential metabolites between CAP patients versus controls, and between the severe CAP group versus non-severe CAP group, were confirmed by targeted mass spectrometry assays in a validation cohort (n = 73, including 32 non-severe CAP, 19 severe CAP and 22 controls). Pearson's correlation analysis was performed to assess relationships between the identified metabolites and clinical severity of CAP. The area under the curve (AUC), sensitivity and specificity of the metabolites for predicting the severity of CAP were also investigated.

Results: The metabolic signature was markedly different between CAP patients and controls. Fifteen metabolites were found to be significantly dysregulated in CAP patients, which were mainly mapped to the metabolic pathways of sphingolipid, arginine, pyruvate and inositol phosphate. The alternation trends of five metabolites among the three groups including sphinganine, p-Cresol sulfate, dehydroepiandrosterone sulfate (DHEA-S), lactate and L-arginine in the validation cohort were consistent with those in the discovery cohort. Significantly lower concentrations of sphinganine, p-Cresol sulfate and DHEA-S were observed in CAP patients than in controls (p < 0.05). Serum lactate and sphinganine levels were positively correlated with confusion, urea level, respiratory rate, blood pressure, and age > 65 years (CURB-65), pneumonia severity index (PSI) and Acute Physiology and Chronic Health Evaluation II (APACHE II) scores, while DHEA-S inversely correlated with the three scoring systems. Combining lactate, sphinganine and DHEA-S as a metabolite panel for discriminating severe CAP from non-severe CAP exhibited a better AUC of 0.911 (95% confidence interval 0.825-0.998) than CURB-65, PSI and APACHE II scores.

Conclusions: This study demonstrates that serum metabolomics approaches based on the LC-MS/MS platform can be applied as a tool to reveal metabolic changes during CAP and establish a metabolite signature related to disease severity.

Trial registration: ClinicalTrials.gov, NCT03093220 . Registered retrospectively on 28 March 2017.

Keywords: Biomarker; Community-acquired pneumonia; Liquid chromatography–mass spectrometry; Metabolomics; Severity.

Conflict of interest statement

Ethics approval and consent to participate

The study protocol was approved by the Institutional Review Board of the Peking University People’s Hospital (Beijing, China). Written informed consent was obtained from all patients or their surrogates.

Competing interests

The authors declare that they have no competing interests.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Figures

Fig. 1

Flowchart of study population enrolment. CAP community-acquired pneumonia, ICU intensive care unit, NSCAP non-severe CAP, SCAP severe CAP

Fig. 2

Multivariate statistical analysis of serum samples in discovery cohort. a PCA score plots. Five samples (four severe CAP and one non-severe CAP) are placed outside the ellipse that describes the 95% CI of Hotelling’s T-squared distribution. b OPLS-DA three-dimensional score plot discriminates all CAP subjects versus controls in discovery cohort (R2Y = 0.937, Q2 = 0.814, p < 0.0001). c OPLS-DA score plots of non-severe CAP versus severe CAP groups (R2Y = 0.757, Q2 = 0.465, p < 0.0001). d OPLS-DA score plots of non-severe CAP patients versus controls (R2Y = 0.994, Q2 = 0.955, p < 0.0001). e OPLS-DA score plots of severe CAP patients versus controls (R2Y = 0.996, Q2 = 0.854, p < 0.0001). R2Y represents goodness of fit, Q2 represents goodness of prediction, p value shows significance level of the model. CAP community-acquired pneumonia, PC principal component, QC quality control

Fig. 3

Fifteen metabolites dysregulated in CAP compared to controls. a Hierarchical cluster heatmap of 15 metabolites in three groups. Row represents metabolites and column represents individual samples. Green, red and blue represent non-severe CAP (NSCAP), severe CAP (SCAP) and controls, respectively. Greater brown indicates higher relative intensity of metabolites, while light blue indicates lower intensity. b Metabolic pathway analysis of 15 metabolites changed in CAP. Node colour based on p value and node radius determined based on pathway impact values. CAP community-acquired pneumonia, DHEA-S dehydroepiandrosterone sulfate

Fig. 4

Three metabolites identified that can discriminate severe CAP from non-severe CAP in validation cohort. Chemical structures of three metabolites. Box–whisker plots of concentrations of three metabolites in three groups. Horizontal line represents median; bottom and the top of box represent 25th and the 75th percentiles; whiskers represent 5% and 95% percentiles. *p < 0.01, **p < 0.001. CAP community-acquired pneumonia, DHEA-S dehydroepiandrosterone sulfate, NSCAP non-severe CAP, SCAP severe CAP

Fig. 5

Correlations of five metabolites with clinical parameters and assessment performance. a Pearson’s correlation heatmap of five serum metabolites and clinical parameters. Greater intensities of brown and blue indicate higher positive or negative correlations, respectively. Resulting correlation matrix presented in Additional file 3: Table S6. b ROC curve analysis of various parameters for discrimination of severe CAP from non-severe CAP. DHEA-S dehydroepiandrosterone sulfate, APACHE II Acute Physiology and Chronic Health Evaluation II, CURB-65 confusion, urea level, respiratory rate, blood pressure, and > 65 years, PSI pneumonia severity index, ESR erythrocyte sedimentation rate, WBC white blood cell, PCT procalcitonin, NE% percentage of neutrophils, CRP C-reactive protein, AUC area under the curve