Advances in metabolomics in critically ill patients with sepsis and septic shock

Swarnima Pandey

doi:10.15441/ceem.24.211

Clin Exp Emerg Med > Volume 12(1); 2025 > Article

Pandey: Advances in metabolomics in critically ill patients with sepsis and septic shock

Review Article

Clin Exp Emerg Med 2025; 12(1): 4-15.

Published online: July 19, 2024

DOI: https://doi.org/10.15441/ceem.24.211

Advances in metabolomics in critically ill patients with sepsis and septic shock

Swarnima Pandey

Department of Pharmaceutical Sciences, University of Maryland School of Pharmacy, Baltimore, MD, USA

Correspondence to: Swarnima Pandey Department of Pharmaceutical Sciences, University of Maryland School of Pharmacy, 20 N. Pine Street, Baltimore, MD 21201, USA Email: pswarnima@rx.umaryland.edu

Received: March 5, 2024 Revised: May 30, 2024 Accepted: June 3, 2024

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/).

Abstract

Sepsis is associated with high morbidity and mortality rates in hospitalized patients. This condition has a complex pathophysiology and can swiftly progress to the severe form of septic shock, which can lead to organ dysfunction, organ failure, and death. Metabolomics has transformed the clinical and research topography of sepsis, with application to prognosis, diagnosis, and risk assessment. Metabolomics involves detecting and analyzing levels of metabolites in blood (plasma, serum, and/or erythrocytes) and urine; when applied in sepsis, this technology can improve our understanding of the pathogenesis of the disease and aid in better disease management by identifying early biomarkers. For this review article, “metabolomics,” “sepsis,” and “septic shock” were keywords used to search records in various databases including PubMed and Scopus from their inception until December 2023. This review article summarizes information regarding metabolic profiling performed in sepsis and septic shock and illustrates how metabolomics is advancing the diagnosis and prognosis of patients with sepsis.

Keywords: Metabolomics; Sepsis; Septic shock; Biomarkers; Diagnosis

Capsule Summary

What is already known

Sepsis is associated with high mortality rates, with early diagnosis being critical for improving outcomes. Traditional diagnostic approaches for sepsis may have limitations.

What is new in the current study

This review comprehensively summarizes metabolomic profiling studies in sepsis and septic shock, highlighting key biomarkers for diagnosis, severity assessment, mortality prediction, and treatment response monitoring. Metabolomics offers promising tools for improving patient outcomes through personalized interventions and better understanding of sepsis pathophysiology.

INTRODUCTION

Sepsis accounts for substantial mortality among patients admitted to the intensive care unit. In addition to its highly adverse health effects, it poses a substantial financial burden due to prolonged hospital stays. Sepsis is characterized by a hyperinflammatory response in response to infection, followed by an immunosuppressive phase during which multiple organs are dysfunctional, and patients with sepsis are highly susceptible to infection. In 2016, the Sepsis-3 Conference defined sepsis as a ''life-threatening organ dysfunction caused by a dysregulated host response to infection'' and septic shock as a ''subset of sepsis in which underlying circulatory and cellular/metabolic abnormalities are profound enough to increase mortality substantially'' [1].

Early detection of sepsis is key to preventing its progression to septic shock, which is associated with a mortality rate of 30% to 70% [2,3]. The golden hours for patient survival are the initial hours postdiagnosis, and patient survival is dependent on aggressive treatment during this period. Children in whom septic shock is recognized early and adequately treated have a much higher survival rate than children diagnosed later [4]. Early diagnosis of sepsis is critical because mortality increases by 7.6% for each hour that appropriate antimicrobial therapy is delayed [5–7]. Thus, diagnostic approaches that accelerate disease recognition are essential to improve patient outcomes and decrease mortality [8,9].

Diagnostic criteria for sepsis are nonspecific. Hence, identifying specific and sensitive biomarkers or biomarker panels will aid in reducing the mortality [10,11].

In 1940, Roger Williams was the first to introduce the concept of a “metabolic fingerprint” as a characteristic trait of every individual [12]. The term “metabolomics” was then introduced to describe the scientific discipline that deals with the identification of metabolites that characterize cellular biological processes. Metabolites are the end products of proteomic and genomic processes and are the ultimate phenotype. Thus, they could be ideal biomarkers for diseases and their progression; in addition, metabolomics can help identify the efficacy of treatment.

In 1984, Nicholson et al. [13] demonstrated that nuclear magnetic resonance (NMR) spectroscopy could be used to diagnose diabetes mellitus. Metabolomics has resulted in the identification of various biomarkers that have improved the risk stratification of diabetes and its complications and has provided novel insights into its diagnosis, prognosis, and therapeutic targets.

Metabolomics provides a holistic view of complex metabolic pathways. Since subtle changes in genes and protein can bring about substantial changes in metabolite levels, analysis of metabolites can provide information about the biological status of an individual. Initial alterations in metabolite levels may predict disease severity, and changes observed over time may help characterize therapeutic response, disease progression, and/or clinical outcomes [14,15]. Differences in metabolite levels and their modifications may be associated with biological aberrations that could provide insight into disease pathogenesis [16–18]. Metabolomics has the potential to provide unique insights into metabolic changes in living systems [19,20].

This review provides a snapshot of metabolic outcomes in sepsis and septic shock based on review of several types of metabolomics studies: (1) animal studies and (2) clinical studies (including the subcategories of critically ill patients with systemic inflammatory response syndrome (SIRS) vs. healthy controls, critically ill patients with sepsis vs. healthy controls, critically ill patients with sepsis vs. noninfected SIRS patients, critically ill patients with sepsis vs. healthy controls vs. SIRS, critically ill patients with sepsis vs. intensive care unit [ICU] controls, mortality markers, and treatment response markers).

METHODS

The following databases were searched for references: PubMed, Web of Science, Cochrane Library, and Scopus, from their inception to December 2023. The following terms were utilized: “metabolomics,” “critically ill,” “sepsis,” ”septic shock,” and “metabolic profiling.” A total of 98 articles were identified, of which 68 were utilized for this review.

Animal studies

Animal models of sepsis and septic shock are described in Table 1 [17,21–27]. Sepsis was induced in rats by cecal ligation and puncture (CLP) or lipopolysaccharide (LPS)-induced endotoxemia [21–24]. Studies using CLP-induced sepsis used plasma for NMR analysis and sham surgery rats as controls [22,24]. There were differences in alanine, acetoacetate, and formate levels in the animal studies mentioned above. Izquierdo-Garcia et al. [23] reported an increase in phosphoethanolamine in the sepsis group, while Lin et al. [22] reported an increase in lactate and ketone bodies in septic rates and described sepsis mortality markers. The discriminatory metabolites identified in the study were generated by anaerobic and fatty acid metabolism aberrations, while the increase in formate levels in sepsis was due to the increased synthesis of nucleic acids.

Researchers have also used liquid chromatography (LC) mass spectrometry (MS)-based analysis of plasma [21,24] and urine [24] samples in septic shock-induced rat models. Liu et al. [21] investigated four groups with differential combination-induced septic shock (CLP-induced sepsis and sham burns, sham sepsis and burns, sepsis and burns, and shams of both procedures). Laiakis et al. [24] had five study groups: one with LPS-induced endotoxemia, three exposed to radiation, and a control group. Although there were differences in study design between these two studies, both identified pyrimidines as markers of sepsis, supporting the role of nucleic acids in sepsis.

Furthermore, these studies reported a decline in uric acid in sepsis. The reduction in uric acid is compatible with a reduction in purine turnover, supporting the increased use of nucleic acids in septic shock. Five metabolites were found to be correlated with sepsis and burns, namely cytosine, adenosine, uracil, uric acid, and lactate. These studies illustrated that energy metabolites significantly affect sepsis and septic shock.

Three additional rat studies compared experimental animal models of sepsis and controls [25–27]. Li et al. [25] performed an LC-MS-based metabolomics study to explore metabolic changes in the lymph and plasma as well as lymphatic proinflammatory changes (Tumor necrosis factor–α, interleukin (IL)-1β, and IL-6) in sepsis. Metabolites in lymph fluid that could differentiate septic shock were creatinine, phenylalanine, choline, and vitamin B3 (all elevated in septic shock), while there was a decline in alanine and dimethylarginine in patients with septic shock. These results support the utilization of lipids, protein, and amino acids as alternatives to glucose as an energy source in septic shock, as reported in the previously mentioned studies.

Metabolic differences between the before- and after-induction samples were correlated to identify the significance of acylcarnitine and saturated FAs. Citrulline and lactate were the most significant discriminating metabolites and were validated in a cohort of horses divided into sick and advanced subjects and compared to healthy control horses. The citrulline concentration was lower in the poor outcome cohort than in the healthy group, indicating citrulline as a marker of acute laminitis with a sensitivity of 83% and specificity of 62%.

Langley et al. [27] performed metabolomic and transcriptomic analyses of plasma and tissue samples of the liver, lung, spleen, and blood on days 1, 2, and 5 after infection of primates with Escherichia coli. A decline in lysophosphatidylcholines and increase in kynurenine, bile acids, and tricarboxylic acid (TCA) cycle intermediates were observed in infected animals. Transcriptomics analysis revealed that pathways associated with FA metabolism, branched-chain amino acid (BCAA) catabolism, and inflammation were altered in sepsis. The study concluded that sepsis nonsurvivors had both metabolic and mitochondrial dysfunction, and that the lung was responsible for systemic metabolic responses. Positive correlations were observed among the TCA cycle, inflammatory response, apoptosis and kynurenine pathways, and nonsurvival. Additionally, there were negative correlations between acylglycerophosphocholine (acyl-GPC) and lysophosphatidylcholine acyltransferase 2 (LPCAT2) and nonsurvival [28].

A regression model was built utilizing lysophosphatidylcholine 1 stearoyl GPC, sulfated bile acid, and isovalerylcarnitine. The area under the receiver operating characteristic curve of the model to differentiate infection from noninfection in this primate cohort was determined. This panel of metabolites was able to diagnose sepsis in two human cohorts: the RoCI (Registry of Critical Illness) cohort [29] and the CAPSOD (Community-Acquired Pneumonia and sepsis Outcome Diagnostic) cohort [30]. FA and amino acid metabolism were found to be correlated with mortality.

Liu et al. [21] examined metabolic aberrations associated with two herbal remedies, LXHX (liangxuehuoxue) and QRJD (qinrejiedu), in a mouse model. They used three CLP-induced sepsis groups and a control group and identified 18 metabolites related to energy metabolism, lipid transport, and amino acid that had significantly altered levels in sepsis.

Another study by Zhang et al. [31] illustrated that lysine supplementation in septic mice resulted in less inflammation and less hypotension than placebo administration.

The findings of the studies discussed above suggest that aberrations in FA and amino acid metabolites can be used as potential biomarkers to diagnose sepsis and predict mortality.

Clinical studies

Critical ill patients with SIRS vs. healthy controls

Metabolic profiling in critically ill patients was first performed in trauma patients [32,33], as shown in Table 2 [10,30,32–66]. These studies compared uninfected SIRS versus multiorgan failure (MOF) patients [32] and survivors versus nonsurvivors of septic shock [33].

SIRS was correlated with increased BCAA and glucose levels according to Mao et al. [32], while MOF patients showed increased creatinine, lactate, and free FAs compared to uninfected SIRS patients. Cohen et al. [33] reported increased lipid levels as well as glucose, ketone body, and lactate levels in nonsurvivors versus survivors of septic shock.

Park et al. [34] investigated the use of albumin to treat acute lung injury (ALI) and reported an improvement in oxygenation in the treatment groups compared to the placebo group. These authors analyzed metabolic profiles on days 1, 2, 3, and 7 and compared metabolite levels between the treatment groups and healthy control group. Statistical analysis failed to illustrate any difference in the metabolic profiles of the two groups initially, but differences were observed from day 2 onward. The study reported metabolic differences between treatment groups, with elevation of albumin on day 2, low-density lipoprotein and alanine on day 3, and cholesterol on days 2 and 3. The influence of time on the concentration of discriminatory metabolites was assessed, and high-density lipoprotein, alanine, and valine levels increased over time in the albumin group. This study provided insight into the role of these metabolites in the pathogenesis of the diseases and established the significance of serial studies in tracking metabolic changes related to multiorgan dysfunction and clinical outcomes.

The predictive ability of the metabolites was also demonstrated on day 7 in the albumin treatment group. Clustering of patients with ALI and other underlying disease conditions from the two treatment groups illustrated that the response to ALI had a larger effect on metabolic profiles than did the etiology of ALI.

Critically ill patients with sepsis vs. healthy controls

Pandey [35] integrated clinical data with metabolomics based on a review of the literature to enhance understanding of the condition of septic patients to enable better stratification and improve prediction of their clinical outcomes. Serum and plasma are the two standard biomaterial samples used for metabolomic studies; however, one study used erythrocytes in addition to plasma [36]. One of the earlier metabolomic studies of sepsis was by Stringer et al. [37]; the authors reported a decline in sphingomyelin and elevation in adenosine, glutathione, and phosphatidylserine levels in ALI compared to healthy controls. The pathways affected were associated with oxidation, apoptosis, and energy utilization. The metabolites related to energy utilization were similar to those reported in previous animal model studies, namely pyruvate, ketone bodies, and FA metabolites.

Bruegel et al. [38] in 2012 performed whole blood-based metabolomics using LC-MS-MS. The study included LPS-activated and nonactivated whole blood samples and identified 7 amino acids, five arachidonic acids, and two cyclooxygenase metabolites that were significantly associated with sepsis. LPS-activated blood samples had smaller increases in the levels of amino acids and cyclooxygenase metabolites than healthy controls. A larger increase in these metabolites between the two groups was associated with favorable clinical outcomes at day 14 and reduced disease severity.

A nontargeted metabolomics study was performed by Liang et al. [39], who identified sphingosine, 5-methylcytidine, and 3-dehydrocarnitine as biomarkers for early diagnosis of septic shock. Jaurila et al. [40] validated previously identified shock biomarkers. They reported elevated levels of creatinine, 3-hydroxybutyrate, glycoprotein, and glycine and a decline in citrate and histidine in septic shock. Pandey et al. [43] illustrated that diabetes and hypertension [41], sex [42], and sepsis progression were associated with characteristic biomarkers that can be assessed in the serum samples of patients with sepsis and septic shock. Recent studies by Li et al. [44] identified biomarkers with the ability to predict sepsis and septic shock.

Critically ill patients with sepsis vs. noninfected SIRS patients

The previously reported studies that performed metabolic profiling of sepsis in clinical settings using NMR and LC-MS-MS were pioneer studies. There have since been several other studies that have compared the metabolomic profile of sepsis with that of controls.

Schmerler et al. [67] used LC MS-based plasma metabolomics and demonstrated metabolic differences between sepsis and noninfected SIRS samples. They reported that acylcarnitine and glycerophosphatidylcholines are discriminatory markers of sepsis.

An NMR-based plasma metabolic analysis by Blaise et al. [46] investigated sepsis in trauma patients and reported elevation of TCA intermediates, BCAA ketone bodies, and allantoin. Allantoin is one of the markers of oxidative stress because, under homeostasis, the end product of purines is uric acid, which is converted by reactive oxygen species to allantoin. Thus, oxidative stress is responsible for increasing allantoin and decreasing uric acid, as shown in the experimental model-based study of Liu et al. [21].

Lipids like glycerolipids and FAs have also been found to be metabolites with discriminatory potential and mortality markers in sepsis and septic shock. Stringer et al. [37] reported that glutathione was a potential biomarker of ALI-induced sepsis.

The CAPSOD cohort was utilized by Kamisoglu et al. [30] in 2015 to identify biomarkers of sepsis. The study illustrated that the LPS-induced endotoxemia and sepsis groups had 16 significantly altered metabolites, whereas 18 metabolites were significantly different between noninfected SIRS and LPS-induced endotoxemia groups. Metabolites common to endotoxemia and sepsis were 2-hydroxybutyrate, mannose, bilirubin, and lipids. Acylcarnitine was identified as a mortality marker.

Upon examining survivors, nonsurvivors, and LPS-induced endotoxemia patients, variations were observed in 19 metabolites between patients with LPS-induced endotoxemia and survivors that could differentiate between positive and negative clinical outcomes. The authors hypothesized that LPS-induced endotoxemia includes an induction stage followed by a recovery stage that mirrors responses required for host survival and is different from the adaptive response of sepsis associated with mortality. This study highlighted the potential of metabolic profiling in distinguishing between infected and noninfected SIRS patients.

A milestone study to identify mortality markers of sepsis was conducted by Langley et al. [10] in 2013. Acylcarnitine was able to distinguish sepsis survivors and nonsurvivors. Comparisons of sepsis survivors and patients with noninfective SIRS revealed lower levels of citrate, malate, amino acids, and carnitine esters and an increase in six acetaminophen catabolites in sepsis survivors.

The GenIMS (Genetic and Inflammatory Markers of Sepsis) cohort was subjected to an LC-MS-based metabolomics study by Seymour et al. [47]. Statistical analysis was performed to identify metabolites of oxidative stress, bile acid, nucleic acid, and stress that could distinguish sepsis survivors and nonsurvivors (90-day survival). Pseudouridine was found to be highly significantly associated with survival.

In conclusion, all these studies explored and reported variations in metabolites in patients with sepsis and noninfected SIRS. A decrease in BCAA and increase in non-BCAA, ketone bodies, and intermediates of TCA were common findings. Nonsurvivors were characterized by decreased glycerophospholipid levels and increases in acylcarnitine levels, nucleic acids, and ketone bodies.

Another study in 2012 compared metabolic profiles based on the type of infection [67]. A comparison was made between SIRS and sepsis, with patients divided into test and confirmation cohorts. Patients were stratified according to the type of infection as patients with SIRS, community-acquired pneumonia, urinary tract infections, intraabdominal disease, and bloodstream infections. The study reported alterations in acylcarnitine, glycerophospholipids, and sphingolipids in sepsis patients compared to SIRS patients. They noted that specific metabolites could be used to distinguish the types of infections. They also analyzed the metabolic profiles of patients with poor and good clinical outcomes and reported variations in metabolites depending upon the underlying type of infection.

A recent study by Feng et al. [48] identified succinic acid semialdehyde, uracil, and uridine as diagnostic biomarkers of sepsis in individuals with multiple traumas.

Critically ill patients with sepsis vs. SIRS patients and healthy controls

A study that performed metabolomic analysis of sepsis and septic shock in a pediatric population based on serum samples from septic shock, SIRS, and healthy controls using a nontargeted metabolomics approach identified 2-hydroxybutyrate, lactate, histidine, phenylalanine, and arginine as discriminatory metabolites for septic shock [4].

In another study that compared patients with sepsis, non-SIRS patients, and healthy controls, statistical decreases in lactitol dehydrate and S-phenyl-d-cysteine and increases in S-(3-methylbutanoyl)-dihydrolipoamide-E and N-non-anoyl glycine were found in sepsis patients compared to noninfected SIRS patients. Metabolites associated with the severity of sepsis and mortality were also examined, and decreases in phospholipids, Ne-dimethyl-lysine, intermediates of phenylalanine metabolism, and cysteine were associated with sepsis severity. In contrast, S-(3-methylbutanoyl)-dihydrolipoamide-E, glycerophosphocholine, and S-succinyl glutathione were identified as mortality markers [49].

Critically ill patients with sepsis vs. ICU controls

Differences in the metabolomic profiles of bacteremic sepsis patients and ICU control patients were investigated by Kauppi et al. [50] using whole blood samples. The study identified six significant metabolites. Of the six, myristic acid was the most significant metabolite predictive of sepsis, with a sensitivity of 1.00 and specificity of 0.95, and showed better performance than various combinations of conventional laboratory and clinical parameters.

Mickiewicz et al. [51,52] demonstrated that ¹H NMR could be used as a diagnostic tool for septic shock. In particular, they reported that a decline in glutamine, glutamate, BCAA, and arginine and an elevation in aromatic amino acids and proline characterized septic shock patients.

Mortality biomarkers of septic shock

Several studies have investigated markers of mortality in sepsis and septic shock. Nonsurvivors of sepsis have been shown to have increased amino acid and ketone levels and decreased levels of FA metabolites [53].

Ferrario et al. [54] designed a study comparing 28- and 90-day mortality groups. Spermidine, putrescine, kynurenine, and glucogenic amino acids were elevated and phosphatidylcholines and lysophosphatidylcholines were decreased in nonsurvivors of septic shock. Tryptophan catabolism and lipids were also found to influence septic shock mortality.

Another targeted metabolomics study performed lipid profiling and investigated the association between lipid profiles and mortality in septic shock [55]. The study demonstrated a significant elevation in prostaglandin F2a and leukotriene B4 in nonsurvivors. The association between acetylcarnitine with sepsis and mortality has also been studied; nonsurvivors at 28 days were found to have elevated plasma acetylcarnitine levels [56].

Rogers et al. [68] designed a study to identify metabolites related with 28-day mortality using two cohorts. They identified gamma-glutamyl phenylalanine, gamma-glutamyl tyrosine, 1-arachidonoylGPC(20:4), taurochenodeoxycholate, 3-(4-hydroxyphenyl) lactate, sucrose, kynurenine [68] as significantly associated with mortality [56].

Liu et al. [58] performed a mortality study in septic shock survivors and nonsurvivors using 0- and 24-hour serum samples. Analysis of serum collected at the time of admission (0 hour) revealed differences in the metabolic profiles of survivors and nonsurvivors. There were increases in creatinine, energy metabolites, and amino acid levels and down-regulation of glycoprotein concentrations over time from 0 to 24 hours in nonsurvivors. Significant metabolites that distinguished septic shock survivors from nonsurvivors over a period of 24 hours were alanine, glutamate, lactate, pyruvate, N-acetyl glycoprotein, and citrate. The study demonstrated that monitoring relevant metabolites can help determine early therapeutic responses.

Twenty-one cohorts comprising 1,287 individuals and 2,509 metabolites were analyzed by Wang et al. [69] in a meta-analysis, and the authors identified specific amino acids, mitochondrial metabolites, eicosanoids, and lysophospholipids as biomarkers of sepsis.

Garcia-Simon et al. [59] used 0- and 24-hour urine samples and identified arginine, methionine, phenylalanine hippurate, and ethanol as markers of mortality. Phenylalanine and leucine were used for risk stratification in patients with sepsis and septic shock by Huang et al. [60]. Other high risk markers, namely symmetrical dimethylarginine and asymmetrical dimethylarginine, were identified in sepsis by Winkler et al. [61].

Cambiaghi et al. [62] reported dynamic changes in metabolite levels over the study period in severe septic shock patients stratified for mortality. Meanwhile, Evans et al. [63] demonstrated decreased phenylalanine levels in septic shock nonsurvivors at 1 year.

Treatment response biomarkers in sepsis and septic shock

Two metabolomic studies have monitored the treatment response in sepsis [64,65]. Puskarich et al. [64] included patients receiving L-carnitine supplementation. Patients with a good response had low levels of carnitine and acetylcarnitine, while methionine, lysine, phenylalanine, and tyrosine levels were increased after L-carnitine was administered. Cambiaghi et al. [65] categorized septic shock patients based upon the Sequential Organ Failure Assessment score as responders or nonresponders. Myristic acid and oleic acid showed a larger decrease while creatinine showed a smaller decrease in responders than nonresponders. Over time, kynurenine increased in responders but not in nonresponders.

Pandey et al. [57] used metabolomics to monitor treatment efficacy during hospital stay for sepsis or septic shock and reported that a metabolomics approach was suitable for this purpose.

LIMITATIONS AND FUTURE PROSPECTIVES OF METABOLOMICS IN SEPSIS AND SEPTIC SHOCK

Although precise biomarkers of sepsis have been identified in metabolomics studies, there needs to be greater standardization, which is one of the significant issues when attempting to replicate results. Moreover, the complexity and heterogeneity of sepsis create a large variety of study populations. Therefore, combining conventional biomarkers and metabolomic profiling is likely to be necessary. Moreover, multiomics should be performed simultaneously given that sepsis is also associated with alterations in protein and gene expression [70,71]. Metabolomics can be utilized to identify mitochondrial dysfunction or aberrations in the microcirculation in sepsis. Future metabolomics studies should aim to develop bedside laboratory kits for clinical practice. Metabolomics studies to predict the outcomes of sepsis and septic shock patients are ongoing [71]. Identifying mortality markers at an early stage should improve patient outcomes. Ideally, a pharmacometabolomics approach should be used to determine the appropriate drugs or to identify patients likely to respond to specific therapies.

Due to the overwhelming amount of metabolomic information associated with sepsis, artificial intelligence and machine learning should be utilized to handle the enormous amount of data.

CONCLUSION

Metabolic profiling studies of sepsis have established that metabolomics has the potential to be used as a diagnostic and prognostic tool capable of providing biomarkers for early diagnosis, prognosis, severity determination, and mortality prediction. Numerous metabolomics-based biomarkers have been identified; by integrating this knowledge in this review, I hope to have advanced our understanding of metabolomics in sepsis and septic shock to provide a better understanding of needed improvements in the management and outcomes of patients with either sepsis or septic shock.

NOTES

Conflicts of interest

The author has no conflicts of interest to declare.

Funding

The author received no financial support for this study.

Data availability

Data sharing is not applicable as no new data were created or analyzed in this study.

REFERENCES

1. Singer M, Deutschman CS, Seymour CW, et al. The third international consensus definitions for sepsis and septic shock (Sepsis-3). JAMA 2016; 315:801-10.

2. Namgung M, Ahn C, Park Y, Kwak IY, Lee J, Won M. Mortality among adult patients with sepsis and septic shock in Korea: a systematic review and meta-analysis. Clin Exp Emerg Med 2023; 10:157-71.

3. Kadri SS, Rhee C, Strich JR, et al. Estimating ten-year trends in septic shock incidence and mortality in United States academic medical centers using clinical data. Chest 2017; 151:278-85.

4. Mickiewicz B, Vogel HJ, Wong HR, Winston BW. Metabolomics as a novel approach for early diagnosis of pediatric septic shock and its mortality. Am J Respir Crit Care Med 2013; 187:967-76.

5. Chatterjee S, Bhattacharya M, Todi SK. Epidemiology of adult-population sepsis in India: a single center 5 year experience. Indian J Crit Care Med 2017; 21:573-7.

6. Paary TT, Kalaiselvan MS, Renuka MK, Arunkumar AS. Clinical profile and outcome of patients with severe sepsis treated in an intensive care unit in India. Ceylon Med J 2016; 61:181-4.

7. Sharma SK, Rohatgi A, Bajaj M, et al. Sepsis 2016 Agra, India. Agra, India. 5-6 February 2016. Crit Care 2016; 20 Suppl 1(Suppl 1):45.

8. Rivers E, Nguyen B, Havstad S, et al. Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med 2001; 345:1368-77.

9. Evangelatos N, Bauer P, Reumann M, Satyamoorthy K, Lehrach H, Brand A. Metabolomics in sepsis and its impact on public health. Public Health Genomics 2017; 20:274-85.

10. Langley RJ, Tsalik EL, van Velkinburgh JC, et al. An integrated clinico-metabolomic model improves prediction of death in sepsis. Sci Transl Med 2013; 5:195ra95.

11. Liu X, Ren H, Peng D. Sepsis biomarkers: an omics perspective. Front Med 2014; 8:58-67.

12. Gates SC, Sweeley CC. Quantitative metabolic profiling based on gas chromatography. Clin Chem 1978; 24:1663-73.

13. Nicholson JK, O'Flynn MP, Sadler PJ, Macleod AF, Juul SM, Sonksen PH. Proton-nuclear-magnetic-resonance studies of serum, plasma and urine from fasting normal and diabetic subjects. Biochem J 1984; 217:365-75.

14. Banoei MM, Donnelly SJ, Mickiewicz B, Weljie A, Vogel HJ, Winston BW. Metabolomics in critical care medicine: a new approach to biomarker discovery. Clin Invest Med 2014; 37:E363-76.

15. Eckerle M, Ambroggio L, Puskarich MA, et al. Metabolomics as a driver in advancing precision medicine in sepsis. Pharmacotherapy 2017; 37:1023-32.

16. Parent BA, Seaton M, Sood RF, et al. Use of metabolomics to trend recovery and therapy after injury in critically ill trauma patients. JAMA Surg 2016; 151:e160853.

17. Steelman SM, Johnson P, Jackson A, Schulze J, Chowdhary BP. Serum metabolomics identifies citrulline as a predictor of adverse outcomes in an equine model of gut-derived sepsis. Physiol Genomics 2014; 46:339-47.

18. Siddiqui MA, Pandey S, Azim A, Sinha N, Siddiqui MH. Metabolomics: an emerging potential approach to decipher critical illnesses. Biophys Chem 2020; 267:106462.

19. Fanos V, Stronati M, Gazzolo D, Corsello G. Metabolomics in the diagnosis of sepsis. Ital J Pediatr 2014; 40(Suppl 1):A11.

20. Pandey S. Metabolomics characterization of disease markers in diabetes and its associated pathologies. Metab Syndr Relat Disord 2024; 22:499-509.

21. Liu XR, Zheng XF, Ji SZ, et al. Metabolomic analysis of thermally injured and/or septic rats. Burns 2010; 36:992-8.

22. Lin ZY, Xu PB, Yan SK, et al. A metabonomic approach to early prognostic evaluation of experimental sepsis by (1)H NMR and pattern recognition. NMR Biomed 2009; 22:601-8.

23. Izquierdo-Garcia JL, Nin N, Ruiz-Cabello J, et al. A metabolomic approach for diagnosis of experimental sepsis. Intensive Care Med 2011; 37:2023-32.

24. Laiakis EC, Hyduke DR, Fornace AJ. Comparison of mouse urinary metabolic profiles after exposure to the inflammatory stressors γ radiation and lipopolysaccharide. Radiat Res 2012; 177:187-99.

25. Li Y, Hou M, Wang JG, et al. Changes of lymph metabolites in a rat model of sepsis induced by cecal ligation and puncture. J Trauma Acute Care Surg 2012; 73:1545-52.

26. Li Y, Liu H, Wu X, Li D, Huang J. An NMR metabolomics investigation of perturbations after treatment with Chinese herbal medicine formula in an experimental model of sepsis. OMICS 2013; 17:252-8.

27. Langley RJ, Tipper JL, Bruse S, et al. Integrative “omic” analysis of experimental bacteremia identifies a metabolic signature that distinguishes human sepsis from systemic inflammatory response syndromes. Am J Respir Crit Care Med 2014; 190:445-55.

28. Yamashita A, Hayashi Y, Nemoto-Sasaki Y, et al. Acyltransferases and transacylases that determine the fatty acid composition of glycerolipids and the metabolism of bioactive lipid mediators in mammalian cells and model organisms. Prog Lipid Res 2014; 53:18-81.

29. Dolinay T, Kim YS, Howrylak J, et al. Inflammasome-regulated cytokines are critical mediators of acute lung injury. Am J Respir Crit Care Med 2012; 185:1225-34.

30. Kamisoglu K, Haimovich B, Calvano SE, et al. Human metabolic response to systemic inflammation: assessment of the concordance between experimental endotoxemia and clinical cases of sepsis/SIRS. Crit Care 2015; 19:71.

31. Zhang Y, Yu W, Han D, Meng J, Wang H, Cao G. L-lysine ameliorates sepsis-induced acute lung injury in a lipopolysaccharide-induced mouse model. Biomed Pharmacother 2019; 118:109307.

32. Mao H, Wang H, Wang B, et al. Systemic metabolic changes of traumatic critically ill patients revealed by an NMR-based metabonomic approach. J Proteome Res 2009; 8:5423-30.

33. Cohen MJ, Serkova NJ, Wiener-Kronish J, Pittet JF, Niemann CU. 1H-NMR-based metabolic signatures of clinical outcomes in trauma patients: beyond lactate and base deficit. J Trauma 2010; 69:31-40.

34. Park Y, Jones DP, Ziegler TR, et al. Metabolic effects of albumin therapy in acute lung injury measured by proton nuclear magnetic resonance spectroscopy of plasma: a pilot study. Crit Care Med 2011; 39:2308-13.

35. Pandey S. Sepsis, management & advances in metabolomics. Nanotheranostics 2024; 8:270-84.

36. Mecatti GC, Fernandes Messias MC, Sant'Anna Paiola RM, et al. Lipidomic profiling of plasma and erythrocytes from septic patients reveals potential biomarker candidates. Biomark Insights 2018; 13:1177271918765137.

37. Stringer KA, Serkova NJ, Karnovsky A, Guire K, Paine R, Standiford TJ. Metabolic consequences of sepsis-induced acute lung injury revealed by plasma ¹H-nuclear magnetic resonance quantitative metabolomics and computational analysis. Am J Physiol Lung Cell Mol Physiol 2011; 300:L4-11.

38. Bruegel M, Ludwig U, Kleinhempel A, et al. Sepsis-associated changes of the arachidonic acid metabolism and their diagnostic potential in septic patients. Crit Care Med 2012; 40:1478-86.

39. Liang Q, Liu H, Xing H, Jiang Y, Zhang AH. UPLC-QTOF/MS based metabolomics reveals metabolic alterations associated with severe sepsis. RSC Adv 2016; 6:43293-8.

40. Jaurila H, Koivukangas V, Koskela M, et al. 1H NMR based metabolomics in human sepsis and healthy serum. Metabolites 2020; 10:70.

41. Pandey S, Adnan Siddiqui M, Azim A, Trigun SK, Sinha N. Serum metabolic profiles of septic shock patients based upon co-morbidities and other underlying conditions. Mol Omics 2021; 17:260-76.

42. Pandey S, Siddiqui MA, Trigun SK, Azim A, Sinha N. Gender-specific association of oxidative stress and immune response in septic shock mortality using NMR-based metabolomics. Mol Omics 2022; 18:143-53.

43. Pandey S, Azim A, Sinha N. Longitudinal NMR based serum metabolomics to track the potential serum biomarkers of septic shock. Nanotheranostics 2023; 7:142-51.

44. Li Y, Wang C, Chen M. Metabolomics-based study of potential biomarkers of sepsis. Sci Rep 2023; 13:585.

45. Neugebauer S, Giamarellos-Bourboulis EJ, Pelekanou A, et al. Metabolite profiles in sepsis: developing prognostic tools based on the type of infection. Crit Care Med 2016; 44:1649-62.

46. Blaise BJ, Gouel-Cheron A, Floccard B, Monneret G, Allaouchiche B. Metabolic phenotyping of traumatized patients reveals a susceptibility to sepsis. Anal Chem 2013; 85:10850-5.

47. Seymour CW, Yende S, Scott MJ, et al. Metabolomics in pneumonia and sepsis: an analysis of the GenIMS cohort study. Intensive Care Med 2013; 39:1423-34.

48. Feng K, Dai W, Liu L, et al. Identification of biomarkers and the mechanisms of multiple trauma complicated with sepsis using metabolomics. Front Public Health 2022; 10:923170.

49. Su H, Chang SS, Han CM, et al. Inflammatory markers in cord blood or maternal serum for early detection of neonatal sepsis: a systemic review and meta-analysis. J Perinatol 2014; 34:268-74.

50. Kauppi AM, Edin A, Ziegler I, et al. Metabolites in blood for prediction of bacteremic sepsis in the emergency room. PLoS One 2016; 11:e0147670.

51. Mickiewicz B, Duggan GE, Winston BW, et al. Metabolic profiling of serum samples by 1H nuclear magnetic resonance spectroscopy as a potential diagnostic approach for septic shock. Crit Care Med 2014; 42:1140-9.

52. Mickiewicz B, Tam P, Jenne CN, et al. Integration of metabolic and inflammatory mediator profiles as a potential prognostic approach for septic shock in the intensive care unit. Crit Care 2015; 19:11.

53. Liu Z, Yin P, Amathieu R, Savarin P, Xu G. Application of LC-MS-based metabolomics method in differentiating septic survivors from non-survivors. Anal Bioanal Chem 2016; 408:7641-9.

54. Ferrario M, Cambiaghi A, Brunelli L, et al. Mortality prediction in patients with severe septic shock: a pilot study using a target metabolomics approach. Sci Rep 2016; 6:20391.

55. Dalli J, Colas RA, Quintana C, et al. Human sepsis eicosanoid and proresolving lipid mediator temporal profiles: correlations with survival and clinical outcomes. Crit Care Med 2017; 45:58-68.

56. Chung KP, Chen GY, Chuang TY, et al. Increased plasma acetylcarnitine in sepsis is associated with multiple organ dysfunction and mortality: a multicenter cohort study. Crit Care Med 2019; 47:210-8.

57. Pandey S, Siddiqui MA, Azim A, Sinha N. Metabolic fingerprint of patients showing responsiveness to treatment of septic shock in intensive care unit. MAGMA 2023; 36:659-69.

58. Liu Z, Triba MN, Amathieu R, et al. Nuclear magnetic resonance-based serum metabolomic analysis reveals different disease evolution profiles between septic shock survivors and non-survivors. Crit Care 2019; 23:169.

59. Garcia-Simon M, Morales JM, Modesto-Alapont V, et al. Prognosis biomarkers of severe sepsis and septic shock by 1H NMR urine metabolomics in the intensive care unit. PLoS One 2015; 10:e0140993.

60. Huang SS, Lin JY, Chen WS, et al. Phenylalanine- and leucine-defined metabolic types identify high mortality risk in patients with severe infection. Int J Infect Dis 2019; 85:143-9.

61. Winkler MS, Nierhaus A, Rosler G, et al. Symmetrical (SDMA) and asymmetrical dimethylarginine (ADMA) in sepsis: high plasma levels as combined risk markers for sepsis survival. Crit Care 2018; 22:216.

62. Cambiaghi A, Diaz R, Martinez JB, et al. An innovative approach for the integration of proteomics and metabolomics data in severe septic shock patients stratified for mortality. Sci Rep 2018; 8:6681.

63. Evans CR, Karnovsky A, Puskarich MA, Michailidis G, Jones AE, Stringer KA. Untargeted metabolomics differentiates l-carnitine treated septic shock 1-year survivors and nonsurvivors. J Proteome Res 2019; 18:2004-11.

64. Puskarich MA, Finkel MA, Karnovsky A, et al. Pharmacometabolomics of l-carnitine treatment response phenotypes in patients with septic shock. Ann Am Thorac Soc 2015; 12:46-56.

65. Cambiaghi A, Pinto BB, Brunelli L, et al. Characterization of a metabolomic profile associated with responsiveness to therapy in the acute phase of septic shock. Sci Rep 2017; 7:9748.

66. Chen Q, Liang X, Wu T, et al. Integrative analysis of metabolomics and proteomics reveals amino acid metabolism disorder in sepsis. J Transl Med 2022; 20:123.

67. Schmerler D, Neugebauer S, Ludewig K, Bremer-Streck S, Brunkhorst FM, Kiehntopf M. Targeted metabolomics for discrimination of systemic inflammatory disorders in critically ill patients. J Lipid Res 2012; 53:1369-75.

68. Rogers AJ, McGeachie M, Baron RM, et al. Metabolomic derangements are associated with mortality in critically ill adult patients. PLoS One 2014; 9:e87538.

69. Wang J, Sun Y, Teng S, Li K. Prediction of sepsis mortality using metabolite biomarkers in the blood: a meta-analysis of death-related pathways and prospective validation. BMC Med 2020; 18:83.

70. Cao Z, Robinson RA. The role of proteomics in understanding biological mechanisms of sepsis. Proteomics Clin Appl 2014; 8:35-52.

71. Maslove DM, Wong HR. Gene expression profiling in sepsis: timing, tissue, and translational considerations. Trends Mol Med 2014; 20:204-13.

Table 1.

Animal metabolic profiling studies of sepsis

Study	Platform	Subject	Metabolite identified
Lin et al. [22] (2009)	¹H NMR	CLP-induced sepsis survivors	↑ Alanine, formate, lactate, acetoacetate, hydroxybutyrate, and acetate
		CLP-induced sepsis nonsurvivors
		Sham controls
Liu et al. [21] (2010)	LC-MS/MS	CLP-induced sepsis	(1) Sepsis and burns single pathology groups:
		CLP-induced sepsis and burns	↓ Uric acid and bile acid
		Burns	(2) Sepsis and dual pathology groups:
		Sham of CLP-induced sepsis and burns	↑ Uracil and nitrotyrosine
			(3) Exclusively in dual pathology group:
			↑ Hypoxanthine; indoxyl sulfate tryptophan; ↑ glucuronic acid and gluconic acid and proline ↑
Izquierdo-Garcia et al. [23] (2011)	¹H NMR	CLP-induced sepsis	↑ Alanine, formate, acetoacetate, creatine and phosphoethanolamine and myoinositol
Izquierdo-Garcia et al. [23] (2011)	¹H NMR	Sham surgical controls
Laiakis et al. [24] (2012)	LC-MS/MS	LPS endotoxin model	(1) LPS endotoxin group:
		Radiation with 3 Gy of γ rays	↑ Cytosine, adenosine and O-propanolylcarnitine
		Radiation with 8 Gy of γ rays	(2) LPS endotoxin and 15 Gy radiation group:
		Radiation with 15 Gy of γ rays	↑ Isethionic acid, 2-hydroxyethane-1-sulfonic acid
		Controls
Li et al. [25] (2012)	LC-MS/MS	CLP-induced sepsis	Sepsis vs. noninfected sham controls:
Li et al. [25] (2012)	LC-MS/MS	Sham surgical controls	↑ Palmitoyl-L-carnitine, creatinine, phenylalanine, isonicotinic acid; choline 5-azacytidine and ↓ 1-O-Hexadecyl-2-lyso-glycero-3-phosphorylcholine, alanine, 4-amino-5-hydroxymethyl-2-methylpyrimidine, asymmetric dimethylarginine
Li et al. [26] (2013)	¹H NMR	CLP-induced sepsis model	Sepsis vs. noninfected surgical sham:
		CLP-induced sepsis model (+herb A)	↑ Isobutyrate, 3-hydroxybutyrate, alanine, acetate, lactate and glucose; ↑ TAGs and FAs, ↓ proline, taurine, valine, isoleucine, arginine, lysine and ↑ threonine; ↓ choline and trimethylamine N-oxide
		CLP-induced sepsis model (+herb B)
		Sham surgical control
Steelman et al. [17] (2014)	LC-MS/MS	Cohort 1:	Preinduction vs. postinduction of acute laminitis:
		Preinduction of acute laminitis	↑ Acylcarnitine and amino acids including alanine, kynurenine, taurine and aromatic amino acids and ↓ 3-hydroxybutyrate, citrulline, cysteine, phosphatidylcholine
		Postinduction of acute laminitis	(1) Combined good and poor outcome groups vs. healthy control: ↓ Citrulline
		Cohort 2:	(2) Citrulline as a predictive marker of the development of acute laminitis or nonsurvival
		Poor outcome	(3) Poor outcome and good outcome vs. healthy control:
		Good outcome	↑ Alanine, valine, glycine, and ↓ serine in the good and poor outcome groups compared to healthy controls.
		Healthy controls	(4) Poor outcome and good outcome vs. healthy controls:
			Glycine differed significantly between the two outcome
Langley et al. [27] (2014)	LC-MS/MS	Inoculation of primates with Escherichia coli	Sepsis nonsurvivors vs. survivors and differentiation of sepsis vs. healthy control:
		Sepsis nonsurvivors	↓ Acyl-GPCs, and↑ kynurenine, bile acids, carnitine, TCA cycle
		Sepsis survivors
		Healthy controls

NMR, nuclear magnetic resonance; CLP, cecal ligation and puncture; LC-MS/MS, liquid chromatography coupled mass spectrometry; LPS, lipopolysaccharide; TAG, triacylglyceride; FA, fatty acid; GPC, glycerophosphocholine; TCA, tricarboxylic acid.

Table 2.

Clinical metabolic profiling studies of sepsis and septic shock

Study	Platform	Subject	Metabolite identified
Mao et al. [32] (2009)	¹H NMR	Severe trauma and SIRS	MODS vs. SIRS and both trauma vs. healthy controls
		Severe trauma and MODS	MODS: ↑ Free FAs, glycerol, creatinine, and lactate
		Healthy controls	SIRS: ↑ Amino acids (predominantly BCAAs) and glucose
Cohen et al. [33] (2010)	¹H NMR	Severe trauma survivors	Survivor vs. nonsurvivors and both trauma vs. healthy control:
		Severe trauma and nonsurvivors	↑ Glucose, glutamate, ketone bodies, lactate, TAGs, mono-unsaturated FAs and glycerophospholipids.
		Healthy controls
Park et al. [34] (2011)	¹H NMR	Acute lung injury treatment	Acute lung injury treatment vs. acute lung injury placebo:
		Acute lung injury placebo	↑ Lysyl moiety of albumin; alanine, LDL, VLDL, valine, cholesterol.
		Healthy controls
Stringer et al. [37] (2011)	¹H NMR	Sepsis-related ALI	Sepsis-related ALI vs. healthy controls:
Stringer et al. [37] (2011)	¹H NMR	Healthy controls	↑ Adenosine, glutathione and phosphatidylserine; ↓ sphingomyelin.
Bruegel et al. [38] (2012)	LC-MS/MS	Sepsis	LPS-activated whole blood vs. healthy controls:
Bruegel et al. [38] (2012)	LC-MS/MS	Healthy controls	↓ AA, PGE2, 11-HETE; TXB2.
Langley et al. [10] (2013)	LC-MS/MS	Sepsis nonsurvivors	Sepsis nonsurvivors vs. survivors:
		Sepsis survivors	↑ 17 Amino acid catabolites, 16 carnitine esters, 11 nucleic acid catabolites, citrate, dihydroxyacetone, malate, pyruvate, 4 free FAs; in addition to ↓ 7 GPCs and GPE, ↑ lactate, and acylcarnitines-carnitines.
		Noninfected SIRS	Sepsis survivors vs. noninfected SIRS:
			↓ Citrate and malate, glycerol, glycerol 3-phosphate, phosphate, 21 amino acids and their catabolites, 12 GPCs and GPE esters, and 6 carnitine esters.
Blaise et al. [46] (2013)	¹H NMR	Trauma + sepsis	Trauma + sepsis vs. trauma – sepsis:
Blaise et al. [46] (2013)	¹H NMR	Trauma – sepsis	↑ Aspartate, citrate, valine, hydroxybutyrate, and allantoin.
Seymour et al. [47] (2013)	LC-MS/MS	Sepsis nonsurvivors	Sepsis nonsurvivors vs. survivors:
Seymour et al. [47] (2013)	LC-MS/MS	Sepsis survivors	↑ Taurochenolate sulfate and glycochenolate sulfate; ↑ cortisol, cortisone, and sulfated hormones allantoin, N1-methyladenosine, N methyladenosine, N2, N2-dimethylguanosine, N6-carbamoylthreonyladenosine, and pseudouridine
Su et al. [49] (2014)	LC-MS/MS	Severe sepsis	Sepsis vs. noninfected SIRS:
		Uncomplicated sepsis	↓ Lactitol dehydrate and S-phenyl-d-cysteine and ↑ in S-(3-methylbutanoyl)-dihydrolipoamide-E and N-non-anoyl glycine
		Noninfected SIRS	Severe sepsis vs. uncomplicated sepsis:
		Healthy controls	↓ Glyceryl-phosphoryl-ethanolamine, Ne, Ne-dimethyl-lysine, phenylacetamide and d cysteine
			Death within 24 hr:
			↓ S-(3-methylbutanoyl)-dihydrolipoamide-E, phosphatidylglycerol, glycerophosphocholine GPC, and S-succinyl glutathione
Kamisoglu et al. [30] (2015)	LC-MS/MS	Sepsis nonsurvivors	Sepsis and healthy control vs. LPS-induced endotoxemia vs. noninfected SIRS:
		Sepsis survivors	↑ 2-Hydroxybutyrate, mannose, bilirubin, and lipids
		LPS-induced endotoxemia in healthy controls	Sepsis survivors vs nonsurvivors:
		Noninfected SIRS	↑ Acylcarnitines-carnitines were the most discriminatory metabolites
Mickiewicz et al. [51] (2014)	¹H NMR	Septic shock	Discriminatory metabolite between septic shock, SIRS, and healthy controls:
		SIRS	2 Hydroxybutyrate, lactate, histidine, phenylalaninephenylalanine, and arginine
		Healthy controls
Mickiewicz et al. [52] (2015)	¹H NMR	Septic shock	Discriminatory metabolite between sepsis survivors and nonsurvivors:
Mickiewicz et al. [52] (2015)	¹H NMR	ICU controls	20 Metabolites differentiated the profiles of survivors and nonsurvivors
Garcia-Simon et al. [59] (2015)	¹H NMR	Septic shock	Discriminatory metabolite between sepsis survivors and nonsurvivors:
Garcia-Simon et al. [59] (2015)	¹H NMR	Septic shock	Arginine, methionine, glutamine, phenylalanine, glucose, ethanol, and hippurate showing differences between nonsurvivors and survivors
Liu et al. [53] (2016)	LC-MS/MS	Septic shock	Discriminatory metabolite:
			43 Significant metabolites varied in their levels when compared between survivors with nonsurvivors
			6 Primary discriminators: valine, leucine, isoleucine, citrulline, carnitine 2:0, and betanin
Ferrario et al. [54] (2016)	LC-MS/MS	Septic shock	Upregulated in nonsurvivors:
			Polyamines, glucogenic amino acids, and kynurenine
			Downregulated in nonsurvivors:
			Phosphatidylcholines and lysophosphatidylcholines
Neugebauer et al. [45] (2016)	LC-MS/MS	SIRS	Discriminatory metabolite:
Neugebauer et al. [45] (2016)	LC-MS/MS	Sepsis	Acylcarnitines, glycerophospholipids and sphingolipids were altered in sepsis compared to SIRS
Cambiaghi et al. [62] (2018)	LC-MS/MS	Septic shock	Discriminatory metabolite:
			Alteration in the lipidome of nonsurvivors was found
			PC aa C42:6, PC aa C40:6, and lyso-PC species
Dalli et al. [55] (2017)	LC-MS/MS	Septic shock	Discriminatory metabolite:
Dalli et al. [55] (2017)	LC-MS/MS	Septic shock	Elevation in the levels prostaglandin F2α, leukotriene B4, resolvin E1 resolvin D5, and 17R-protectin D1 were found in nonsurvivors
Chung et al. [56] (2019)	UHPLC-MS	Septic shock	Discriminatory metabolite:
Chung et al. [56] (2019)	UHPLC-MS	Septic shock	A significantly higher level of plasma acetylcarnitine was found in sepsis nonsurvivors when compared with survivors
Liu et al. [58] (2019)	¹H NMR	Septic shock	Discriminatory metabolite:
			The concentrations of alanine, glutamate, glutamine, methionine, aromatic amino acids, ketone bodies, 3-hydroxybutyrate, and acetate were increased in the nonsurvivors as compared to the survivors
			N-acetyl glycoprotein level was found decreased in nonsurvivors
Jaurila et al. [40] (2020)	¹H NMR	Sepsis	Discriminatory metabolite:
Jaurila et al. [40] (2020)	¹H NMR	Healthy controls	Significantly higher serum lactate and citrate concentrations in nonsurvivors compared with survivors
Liang et al. [39] (2016)	LC-MS/MS	Septic shock	Discriminatory metabolite:
Liang et al. [39] (2016)	LC-MS/MS	Healthy controls	Sphingosine, 5 methylcytidine, 3 dehydrocarnitine, 4 acetamido-2-aminobutanoic acid and phenyllactic acid in the septic shock subjects were significantly different from the controls
Feng et al. [48] (2022)	LC-MS/MS	Multiple trauma (non-SIRS)	Discriminatory metabolite:
Feng et al. [48] (2022)	LC-MS/MS	Sepsis	9 Potential biomarkers, namely, acrylic acid, 5-amino-3-oxohexanoate, 3b-hydroxy-5-cholenoic acid, cytidine, succinic acid semialdehyde, PE [P-18:1(9Z)/16:1(9Z)], sphinganine, uracil, and uridine were identified
Mecatti et al. [36] (2018)	LC-MS/MS	Septic shock	Discriminatory metabolite:
		Healthy controls	FAs and phospholipids detected in plasma and erythrocytes could signal sepsis vs. nonsepsis
			Lyso-PCs and SMs were downregulated, whereas the saturated and unsaturated PCs were upregulated in the plasma and erythrocytes of septic patients
Pandey et al. [41] (2021)	¹H NMR	Septic shock	Discriminatory metabolite:
		Healthy controls	The potential biomarkerspotential biomarkers for septic shock are lactate, 3 hydroxybutyrate, 3 hydroxyisovalerate, proline, 1,2 propanediol, creatine, glycine, phenylalanine, and myoinositol, bile, NAG, and VLDL, which were significantly upregulated in septic shock patients, whereas citrate, carnitine, HDL, LDL, and lipoprotein with phosphocholine head group were downregulated in septic shock patients
		Septic shock with comorbid conditions (diabetes and hypertension)
		Septic shock with primary diagnosis (respiratory illness and encephalopathy)
Pandey et al. [57] (2023)	¹HNMR	Septic shock pretreatment	Discriminatory metabolite:
Pandey et al. [57] (2023)	¹HNMR	Septic shock posttreatment	The study showed time-dependent metabolite alteration in ketone bodies, amino acids, choline, and NAG in patients undergoing treatment.
Cambiaghi et al. [65] (2017)	LC- MS/MS	Treatment response in septic shock	Discriminatory metabolite:
			Lipidome alterations play an important role in individual patients' responses to infection
			Furthermore, alanine indicates a possible alteration in the glucose-alanine cycle in the liver, providing a different picture of liver functionality from bilirubin
Kauppi et al. [50] (2016)	LC-MS/MS	Sepsis	Discriminatory metabolite:
Kauppi et al. [50] (2016)	LC-MS/MS	Healthy controls	6 Metabolites were identified for bacteremic sepsis
Winkler et al. [61] (2018)	LC- MS/MS	Sepsis	Discriminatory metabolite:
Winkler et al. [61] (2018)	LC- MS/MS	Sepsis	SDMA and ADMA associated with sepsis mortality
Huang et al. [60] (2019)	LC- MS/MS	Septic shock	Discriminatory metabolite:
Huang et al. [60] (2019)	LC- MS/MS	Septic shock	Phenylalanine- and leucine-defined risk classifications provide metabolic information with prognostic value for patients with severe infection
Li et al. [44] (2023)	LC-MS/MS	Sepsis	Discriminatory metabolite:
Li et al. [44] (2023)	LC-MS/MS	Healthy controls	3-Phenyl lactic acid, N-phenylacetylglutamine, phenylethylamine, traumatin, xanthine, methyl jasmonate, indole, l-tryptophan
Chen et al. [66] (2022)	LC-MS/MS	Sepsis	Discriminatory metabolite:
Chen et al. [66] (2022)	LC-MS/MS	Healthy controls	73 Differentially expressed metabolites that could predict sepsis were identified.
Pandey et al. [42] (2022)	¹H NMR	Septic shock survivor (male/female)	Discriminatory metabolite:
		Septic shock nonsurvivor (male/female)	The energy-related metabolites, ketone bodies, choline, and NAG were found to be primarily responsible for differentiating survivors and nonsurvivors
			The sex-based mortality stratification identified a female-specific association of the anti-inflammatory response, innate immune response, and β oxidation, and a male-specific association of the proinflammatory response to septic shock
Puskarich et al. [64] (2015)	LC- MS/MS	Carnitine treatment response in septic shock	Drug responsive metabolite:
Puskarich et al. [64] (2015)	LC- MS/MS	Carnitine treatment response in septic shock	Responsive towards carnitine treatment
Evans et al. [63] (2019)	LC- MS/MS	Carnitine treatment response in septic shock	Drug responsive metabolite:
Evans et al. [63] (2019)	LC- MS/MS	Carnitine treatment response in septic shock	Metabolic signature of L-carnitine-treated nonsurvivors is associated with a severity of illness (e.g., vascular inflammation) that is not routinely clinically detected
Pandey et al. [43] (2023)	¹H NMR	Treatment response in septic shock	Drug responsive metabolite:
Pandey et al. [43] (2023)	¹H NMR	Treatment response in septic shock	3 Hydroxybutyrate, lactate, and phenylalanine which were lower, whereas glutamate and choline higher in patients showing responsiveness

NMR, nuclear magnetic resonance; SIRS, systemic inflammatory response syndrome; MODS, multiorgan dysfunction syndrome; FA, fatty acid; BCAA, branched-chain amino acid; TAG, triacylglyceride; LDL, low-density lipoprotein; VLDL, very low-density lipoprotein; ALI, acute liver injury; LC-MS/MS, liquid chromatography coupled mass spectrometry; LPS, lipopolysaccharides; AA, arachidonic acid; PGE2, prostaglandin E2; 11-HETE, 11-hydroxyeicosatetraenoic acid; TXB2, thromboxane B2; GPC, glycerophosphocholine; GPE, glycerophosphoethanolamine; ICU, intensive care unit; PC, phosphatidylcholine; UHPLC-MS, ultrahigh performance liquid chromatography coupled with mass spectrometry; SM, sphingomyelin; NAG, N-acetylglycoprotein; HDL, high-density lipoprotein; SDMA, symmetric dimethylarginine; ADMA, assymmetric dimethylarginine.

Adapted from Pandey [35], available under the Creative Commons License.