Unveiling promising sepsis biomarkers: a clinical perspective

Abstract

Sepsis is defined as life-threatening organ dysfunction caused by a dysregulated host response to infection. It is characterized by a high incidence, high mortality rate, and poor prognosis. With the increasing consumption of antibiotics and the growing prevalence of antimicrobial resistance, clinicians are in greater need of easily accessible and effective inflammatory markers to monitor the infection process, guide treatment, and assess patient prognosis. Numerous biomarkers have been studied for their potential value in the diagnosis of sepsis, and more than 250 different biomarkers have been proposed. However, many of these biomarkers are more useful for prognostic evaluation than for diagnosis. Biomarkers such as procalcitonin and C-reactive protein have been extensively studied, with procalcitonin being the only sepsis biomarker included in clinical guidelines. It can be used to guide antibiotic use but should not be used as a standalone diagnostic indicator. C-reactive protein has poor sensitivity and specificity for sepsis diagnosis. In recent years, new biomarkers, such as neutrophil extracellular traps, interleukin-6, histones, cell-free DNA and composite indicators, such as the systemic immune inflammation index,have been identified. This article provides a comprehensive review of the diagnostic value of traditional and recently popular novel biomarkers for sepsis.

KEYWORDS:

• Sepsis
• biomarkers
• procalcitonin
• neutrophil extracellular traps
• cfDNA systemic immune inflammation index

Introduction

Sepsis is a complex syndrome of physiological, pathological, and biochemical changes caused by infection. It is considered an overwhelming, systemic proinflammatory response to infection, followed by an immunosuppressive stage characterized by lymphocyte depletion and secondary infections (Evans et al. 2021). In the early stage, the main manifestations are inflammatory responses caused by the release of cytokines and neutrophils. These cytokines include tumor necrosis factor-α, interleukin-1, interleukin-2, interleukin-6, and interleukin-8.Neutrophil-endothelial cell adhesion, complement activation, and coagulation cascades are activated. Extensive tissue factor expression, fibrin deposition, and impaired anticoagulation mechanisms (including activated protein C) can lead to disseminated intravascular coagulation (DIC), resulting in increased organ dysfunction, bleeding, and mortality (Levi 2001). The diagnostic criteria for sepsis have evolved from Sepsis 1.0 to Sepsis 3.0. Currently, the diagnostic criteria for sepsis in the ICU are as follows: (1) confirmed infection or suspected infection, (2) an increase in the baseline Sequential Organ Failure Assessment(SOFA) score of ≥2 points (Singer et al. 2016). The criteria for diagnosis in the emergency department included the following: (1) suspected infection in patients, (2) qSOFA(quick SOFA) score ≥2 points, (3) SOFA = 1, and (4) an NEWs score of 4–6 points, meeting any of ① + ② - ④ (Bisarya et al. 2022; Roggen et al. 2023). The SOFA score requires time-consuming measurements, and many SOFA scores are not routinely measured or available, leading to poor early diagnostic efficiency. qSOFA calculations are based solely on physiological parameters rather than severity scoring, and the predictive efficacy of qSOFA varies depending on the severity of the disease, measurement time, and method. A large retrospective study in Taiwan demonstrated that the correlation between qSOFA and mortality at 72 h was greater than that at 28 days (Chen et al. 2019). A retrospective cohort study revealed poor predictive ability of qSOFA scores for 28-day mortality, with an AUROC of 0.580 (95% CI 0.550 ∼ 0.620) (Hwang et al. 2018), and the predictive efficacy of qSOFA varies regionally, with relatively high efficacy in developed countries (Li et al. 2024). Therefore, the qSOFA serves only as an early warning system with low sensitivity and is not recommended for use alone as a single screening tool for sepsis or septic shock. Therefore, neither the SOFA nor the qSOFA is sufficient for identifying potential sepsis patients. In 2020, the Lancet reported that globally, there are more than 48.9 million new cases of sepsis annually, with approximately 11 million deaths, accounting for nearly 20% of the total global deaths (Rudd et al. 2020). A meta-analysis by Fleischmann (Fleischmann-Struzek et al. 2020) in 2020 revealed an in-hospital incidence rate of sepsis worldwide of 189/100,000 person-years, with a mortality rate of approximately 26.7%. Five domestic studies reported sepsis incidence rates in the ICU ranging from 20.6% to 50.8%. The overall complication rate of sepsis is estimated to be 33.6% (Liu et al. 2022). Sepsis is one of the most expensive diseases to treat, with an estimated annual healthcare burden of $24 billion. Sepsis survivors have a 15% mortality rate in the first year after discharge and a 6–8% mortality rate in the following 5 years (Liu et al. 2022). Although blood cultures are the gold standard for diagnosing serious infections, only 30% of blood cultures are positive in patients with sepsis (Sweeney et al. 2019). In summary, sepsis is not an isolated disease; its occurrence and progression involve multiple processes, including systemic inflammatory network effects caused by infectious pathogens and their toxins, immune suppression, coagulation dysfunction, and endothelial barrier damage. It is closely associated with pathophysiological changes such as multisystem dysfunction, multiple organ dysfunction, and failure in the host. There is no single biomarker that can definitively diagnose sepsis. Therefore, current research suggests that combined indicators are more effective for theearly identification of sepsis patients and monitoring the occurrence and development of sepsis (Karagoz and Yoldas 2019; Buonacera et al. 2022). Additionally, a growing number of biomarkers are being studied from a pathophysiological perspective in sepsis patients. Hence, this article provides a review of the diagnostic value of the aforementioned traditional and recently popular novel biomarkers for sepsis (Table 2).

Traditional serum markers

C-reactive protein (CRP)

C-reactive protein (CRP) is an acute-phase protein released by liver cells. It is coactivated with the complement system in response to antigens and other stimuli, making it a widely available, cost-effective, and highly sensitive marker of inflammatory states. In healthy individuals, the CRP concentration is extremely low, typically less than 10 mg/L (Li et al. 2014). CRP levels upon admission cannot predict the prognosis of sepsis and respiratory infections (Aulin et al. 2021). However, monitoring CRP levels in the intensive care unit (ICU) on the third day, when the level is greater than 100 mg/L, may be a useful indicator for predicting mortality, similar to the Sequential Organ Failure Assessment (SOFA) score (Ryu et al. 2015). When the body experiences acute trauma and infection, the CRP concentration can rapidly increase, generally starting to rise within 6–12 h of infection onset and rapidly peaking within 24–48 h. This increase often precedes the appearance of clinical symptoms such as fever. However, when the infection is under control, CRP levels can quickly decrease, and these results are less affected by other medications. CRP is helpful for early disease diagnosis, but in addition to bacterial infections, some viral infections, cardiovascular diseases, malignancies, tissue injuries, and noninflammatory conditions can also lead to elevated CRP. Slower changes in CRP levels have been associated with ongoing infections, organ failure, and mortality in the ICU (Moreno et al. 2010). Research indicates that CRP concentrations are unable to predict the onset of late-onset neonatal sepsis (LONS) (Kurul et al. 2021), resulting in decreased specificity for the diagnosis of sepsis-related infections.

Procalcitonin (PCT)

PCT is the precursor substance of the serum calcium-lowering hormone calcitonin. Under normal physiological conditions, PCT is primarily secreted by thyroid C cells. In healthy individuals, PCT levels in plasma are extremely low, generally less than 0.05 ng/mL (Gilbert 2010). PCT has a biological half-life of 22–26 h. PCT is selectively responsive to systemic bacterial, fungal, and parasitic infections. Interferon-γ inhibits PCT production, and PCT does not respond to aseptic inflammation or viral infections or exhibits only a mild response (Watanabe et al. 2016). The presence of lipopolysaccharides in the cell wall of gram-negative bacteria leads to the release of proinflammatory cytokines, which induce the production of PCT. Therefore, PCT levels are significantly greater in patients with gram-negative bacteria than in those with gram-positive bacteria. The optimal cutoff value of PCT for distinguishing between gram-negative and fungal infections was 1.6 ng/mL. However, the accuracy for distinguishing gram-positive and fungal infections is lower. A PCT cutoff value >1.3 ng/mL, helps in ruling out fungal bloodstream infections (Liu et al. 2016). Related studies have confirmed this point, as the cytokine response pattern induced by fungi differs from that caused by bacterial sepsis. Therefore, the diagnostic value of PCT in distinguishing fungal infections from other noninfectious conditions is limited (Dou et al. 2013). Other studies have also indicated that serum procalcitonin levels are greater in patients with gram-negative bacteremia than in those with gram-positive bacteremia or candidemia. This is thought to be due to the lack of cell walls in gram-positive bacteria and the release of large amounts of endotoxins by gram-negative bacteria (Leli et al. 2015; Thomas-Rüddel et al. 2018). Continuously monitoring PCT levels can serve as a reliable indicator of treatment effectiveness in determining the use of antibiotics and guiding the duration of antibiotic therapy (Kyriazopoulou et al. 2021). PCT, a response to hypercalcemia resulting from tissue damage, can indicate that damage has begun to develop, which clinical physicians may not detect (Dolin et al. 2018) Research has shown that PCT is more effective than CRP in diagnosing sepsis. However, noninfectious causes of systemic inflammation, such as shock, trauma, surgery, burns, tumors, autoimmune diseases, medications, and some nonbacterial infections,can also induce the synthesis of procalcitonin. However, the association between these inducers and PCT is not as closely related as that with bacterial infections (Becker et al. 2008). Moreover, in patients with controlled local infections, procalcitonin levels may not increase. Additionally, false-negative results can occur when detecting procalcitonin in the very early stages of infection (Christensen et al. 2014). Considering that noninfectious factors can also lead to elevated PCT levels, PCT is the only sepsis biomarker included in clinical guidelines. It can be used to guide antibiotic therapy but should not be used as a standalone diagnostic indicator.

CRP and PCT have been widely used in the diagnosis of sepsis, and dynamic monitoring of PCT levels can serve as a reliable indicator of the efficacy of antibiotic treatment and guide the duration of antibiotic use. Despite their limited ability to differentiate sepsis from other inflammatory conditions or predict patient prognosis, there is a need for more specific and sensitive biomarkers to improve the diagnosis and treatment of sepsis.

Novel serum markers

Interleukin-6 (IL-6)

Interleukin-6 (IL-6) is a cytokine produced by immune and stromal cells that mediates various biological activities critical for both innate and adaptive immunity. It is rapidly induced and released when the body experiences infection, trauma, surgical procedures, stress responses, brain tumors, and a range of acute inflammatory reactions, making it a sensitive marker for the early diagnosis of sepsis patients (Molano Franco et al. 2019). An IL-6 level more than tenfold greater than the baseline value is considered an early indicator of sepsis, with an increase in the IL-6 level before an increase in the C-reactive protein (CRP) (Dolin et al. 2018). In healthy individuals, the plasma concentration of IL-6 typically ranges from 0.2 to 7.8 pg/mL. In contrast, IL-6 concentrations in adult sepsis patients can exceed 1600 pg/mL and should be measured before antibiotic therapy, with a positive threshold ranging from 40 to 20000 pg/mL (Thompson et al. 2012). Research conducted by Nicole et al^.(Fink-Neuboeck et al. 2016) suggests that IL-6 levels exhibit a predictive peak at 6–8 h post surgery, well in advance of clinical manifestations of postoperative systemic inflammatory response syndrome (SIRS). In comparison, plasma PCT levels reach their peak on the first day after surgery, albeit at a time delay, and CRP levels peak 48 h after surgery. These findings indicate that in patients with postoperative systemic inflammatory response syndrome, IL-6 is the most predictive biomarker, while PCT has increased sensitivity and specificity for differentiation. IL-6 also serves as an early warning indicator for neonatal sepsis (Watanabe et al. 2016). The literature suggests that interleukin-6, interleukin-1, and tumor necrosis factor (TNF) can act as triggers for disseminated intravascular coagulation (DIC) and affect coagulation in sepsis patients (Hoppensteadt et al. 2015). However, in addition to bacterial factors, noninfectious factors, such as tumors (Španko et al. 2021), depression (Ting et al. 2020), surgery (Hu et al. 2018), obstructive sleep apnea syndrome (Imani et al. 2020), pulmonary hypertension (Xu et al. 2023), and autoimmune diseases (Sanchis et al. 2020), can also lead to nonspecific elevation of IL-6, possibly resulting in lower diagnostic specificity for sepsis compared with PCT.

Neutrophil extracellular traps (NETs)

In the initial stages of sepsis, excessive immune system activation and cascading inflammation often accompany immunosuppression (Zhu et al. 2022). Cascadingly, neutrophils reach the site of inflammation, leading to the activation of specific effector functions, such as the release of reactive oxygen species, degranulation, the formation of neutrophil extracellular traps (NETs), and phagocytic activity. NETs are formed through a tightly regulated process of cell death known as NETosis. NETosis involves the activation of neutrophils, which release extracellular nets composed of decondensed chromatin and intracellular granule proteins to capture and kill pathogens. These NETs and their components can damage endothelial cells (Scozzi et al. 2022). NETs are composed of linear, less condensed DNA fibers, along with intracellular proteins from various neutrophil organelles, including histones, neutrophil elastase (NE), myeloperoxidase (MPO), and other proteins. During their generation, NETs can produce and release tissue factors, abnormally activate the coagulation pathway, trigger coagulation cascades, promote platelet activation, and induce platelet aggregation, leading to thrombus formation (Zhou et al. 2022). NETs are mesh-like structures composed of DNA, histones, and antimicrobial proteins, primarily with DNA as the backbone, embedded with histones, neutrophil elastase (NE), myeloperoxidase (MPO), antimicrobial peptides (LL-37), and serine proteases. Experimental detection revolves around its components (Table 1). (1) DNA detection methods: (1.1) Fluorescence staining method (White et al. 2017), (1.2) Immunohistochemistry and confocal microscopy: In the immunohistochemistry process, chromogenic reagents label specific antibodies, and laser confocal microscopy tracks and scans live cell tissues or sections, obtaining high-resolution optical sections of NETs (Chatfield et al. 2018). (2) Protease detection methods: (2.1) detection of neutrophil elastase (NE): (2.1.1) Western blotting: The main process includes electrophoresis, membrane transfer, and immunolabeling. Papayannopoulos et al. analyzed blotting results, determined the purity of NETs components and indentified a new mechanism by which NE can promote chromatin condensation (Papayannopoulos et al. 2010). (2.1.2) Microplate colorimetric method: This method uses microplates as carriers,based on the specific reaction between the sample and the substrate. Karandashova (Karandashova et al. 2018) et al. used an amberyl-Ala-Ala-Pro-Val-p-nitroaniline substrate for microplate titration to more accurately determine the NE concentration. (2.2) Detection of myeloperoxidase (MPO): (2.2.2) Enzyme-linked immunosorbent assay (ELISA): ELISA is commonly used in clinical practice to detect the expression of antibodies, antigens, hormones, and microorganisms in peripheral blood serum. Nizam (Nizam et al. 2014) et al. measured the saliva and serum concentrations of MPO and NE in patients with periodontitis and healthy individuals,and determined the biochemical data of systemic chronic periodontitis patients. (2.2.3) Fluorescence immunoassay: This method, similar to ELISA, relies on specific detection of labeled antibodies, enabling qualitative and quantitative analysis of the sample. Funchal (Tripp et al. 2015) et al. detected NETs induced by fungi, bacteria, and viruses, especially the F protein of respiratory syncytial virus, using a fluorescence immunoassay. Professor Zhang Jiang (Zhang et al. 2022) from Fudan University used a reagent kit to detect cfDNA and MPO-DNA levels in serum culture supernatant,and visualized NETs using immunofluorescence confocal microscopy, finding significantly increased NETs levels in sepsis-associated lung injury. In summary, MPO, DNA, and other components are often detected using ELISA, while immunofluorescence staining is employed to observe NETs. Qualitative detection can also be achieved using transmission electron microscopy and scanning electron microscopy. NETs can induce macrophage pyroptosis, exacerbating the inflammatory response in sepsis. Furthermore, NETs can induce M1-type polarization of lung tissue macrophages, increasing lung inflammation and lung injury (Watanabe et al. 2016). Research has shown an association between NETs and the pathogenesis and progression of sepsis. NETs are upregulated in the inflamed intestinal mucosa, feces, or blood, and they are more prone to infiltrate and damage the intestinal epithelium, leading to intestinal injury. Furthermore, NETs and their associated molecules can directly induce the death of intestinal epithelial cells, and the severity of inflammatory bowel disease is positively correlated with NETs (Chen et al. 2022). NETs exacerbate the inflammatory response in sepsis, causing multiorgan dysfunction. Currently, collaborative efforts are being made to develop therapies targeting NETs for various diseases, and these therapies have shown significant success in preclinical models. These approaches include dissolving NET scaffolds, such as dexamethasone, which has been proven to reduce NET formation and has been shown to be beneficial for COVID-19 patients (Group et al. 2021); administering anti-inflammatory and anticytokine treatments which reduce neutrophil increases, NET formation, and NET-induced thrombosis; activating the IL-6 signaling pathway, which promotes NET formation and lung inflammation; and administering tocilizumab, which has been associated with reduced NET formation. Inhibiting NET formation can lead to a decrease in systemic levels of IL-6 and improved survival rates (Keir and Chalmers 2021). In a lipopolysaccharide-induced septic shock mouse model, it was suggested that sivelestat reduces NET formation by inhibiting NE, thereby improving clinical symptoms of lung injury and increasing survival in septic shock mice (Okeke et al. 2020). Boufenzer (Boufenzer et al. 2021) found in animal experiments that myeloid cell-1 inhibitors can reduce NETosis during septic shock, improve vascular function, and reduce sepsis mortality. Given the abundance of NETs in sepsis, neutralizing NET components may be a useful strategy for improving the outcome of sepsis.

Table 1. Methods for detecting NETs.

Components of NETs	Detection Methods
DNA	Fluorescence staining method
	Immunohistochemistry and confocal microscopy
Neutrophil elastase (NE)	Western blotting method
	Microplate colorimetric method
Myeloperoxidase (MPO)	Enzyme-linked immunosorbent assay (ELISA)
	Flow cytometry
	Fluorescent immunochromatography

Table 2. Differences in relevant sepsis biomarkers.

Biomarker	Predictive value	Diagnostic value	Significance	Current progress in treatment research
CRP	No	Yes	Diagnose sepsis: predict mortality	Widely available and cost-effective inflammatory markers
PCT	Yes	Yes	Diagnose sepsis: Guidance on antibiotic treatment duration	Establishing a procalcitonin research model: Guidance on antibiotic use time (Kyriazopoulou et al. 2021)
IL-6	Yes	Yes	Diagnose sepsis: Diagnosing neonatal sepsis: Predicting the impact of inflammation on renal function and renal drug clearance	IL-6 antagonist tocilizumab for the treatment of sepsis (Hamilton et al. 2023)
NETs	Yes	Yes	Diagnose sepsis: Diagnose ARDS; Diagnose inflammatory bowel disease; Predict organ failure	Dexamethasone reduces the production of NETs (Group et al. 2021): IL-6 antagonist tocilizumab reduces NETs production (Keir and Chalmers 2021): Sivelestat inhibits NE and reduces the formation of NETs (Okeke et al. 2020): Myeloid cell-1 inhibitory drugs reduce NETosis and reduce NETs production (Molano Franco et al. 2019)
Histone	Yes	Yes	Diagnose sepsis: Predict thrombosis and DIC; Predict organ failure; Guide antibiotic treatment	Heparin and APC can reduce histone toxicity (Cheng et al. 2019): Polysialic acid (PSA) antagonizes histone activity (Zlatina et al. 2018)
cfDNA	Yes	Yes	Diagnose sepsis Guidance on antibiotic treatment duration Diagnose ARDS Diagnose infective endocarditis	Effectively clear cf-mtDNA through deoxyribonuclease I (DNase-I) and reduce lung damage (Qiao et al. 2022)
ADM	Yes	Yes	Diagnose sepsis	The nonneutralizing anti-ADM antibody Adrecizumab (Karakas et al. 2020)
			predict mortality

Note: CRP-C-reactive protein; PCT-procalcitonin; IL-6-interleukin 6; NETs- neutrophil extracellular traps; Cell-free DNA-cfDNA; Adrenomedullin-ADM

Histones

Histones are essential components of NETs and are positively charged nuclear proteins involved in packaging DNA into chromatin and regulating gene expression. Under physiological conditions, histones construct nucleosomes and participate in DNA transcription, replication, and repair (Chen et al. 2014). Like many innate immune mediators, histones can trigger inflammatory responses, endothelial cell damage, and cascading coagulation activation (Li et al. 2021). In pathological conditions such as sepsis, trauma, or pancreatitis, histones are translocated from the cell nucleus to the bloodstream, where they act as endogenous danger signals or DAMPs. They can directly induce cytotoxicity in vascular endothelial cells, leading to a significant decrease in cellular vitality (Silk et al. 2017). Histones are associated with edema, increased alveolar wall thickness, and occasional bleeding (Scozzi et al. 2022). Extensive cellular damage can cause histones to be released into the circulation, and their procoagulant, endothelial cell-killing, platelet-activating, and antifibrinolytic capabilities contribute to thrombus formation and DIC in sepsis, further leading to multiorgan failure (Alhamdi and Toh 2017). Histones can induce platelet aggregation and thrombocytopenia but can be inhibited by drugs such as heparin (Alhamdi et al. 2016). In a mouse model of LPS-induced shock, blood histone components were detected three minutes after LPS injury and remained elevated for 24 h, which was significantly earlier than the response of PCT, and these findings could guide antibiotic treatment for sepsis patients (Pan et al. 2017)^. Research has shown that CRP can mitigate histone-mediated toxicity by binding to histones, competing with phospholipid liposomes, preventing histones from integrating into cell membranes, and blocking calcium influx (Abrams et al. 2013). Heparin can effectively bind to positively charged histones, forming noncytotoxic heparin-histone complexes that reduce histone-induced endothelial damage and coagulation activation,alleviating histone toxicity (Zhu et al. 2019). APC is a multifunctional serine protease that significantly reduces histone H3 and H4 levels, effectively blocking histone cytotoxicity (Cheng et al. 2019). Deng et al. (2019) developed a novel anti-H3Cit (citrullinated histone H3) monoclonal antibody [H3Cit mAb] that decreases the levels of cytokines such as IL-1b and TNF-α in the serum, thereby increasing the survival rate of LPS-induced endotoxemic mice. Polysialic acid (PSA) is a promising new candidate drug for preventing excessive NETosis, because it can counteract the cytotoxicity of histones (Zlatina et al. 2018). Although a gold standard method for measuring histone concentrations in sepsis patients has not yet been established, histones are currently quantified primarily through western blotting and enzyme-linked immunosorbent assay (ELISA) (Alhamdi and Toh 2017). Given the crucial role of histones in the development of sepsis, which is involved in virtually every stage of sepsis, there is significant potential for the development of novel histone inhibitors. Histones could serve as diagnostic or prognostic biomarkers for sepsis and therapeutic targets.

Cell-free DNA (cfDNA)

Cell-free DNA (cfDNA) molecules in the plasma are nonrandomly fragmented and carry a wealth of information related to the original tissue (Wang et al. 2023). In sepsis, a dysregulated immune inflammatory response is often triggered and driven by the excessive activation of pattern recognition receptors. Among these danger signals, cell-free DNA (cf-DNA), including nuclear DNA and mitochondrial DNA released from damaged cells, as well as neutrophil extracellular traps, is considered a prognostic and predictive biomarker for sepsis (Huang et al. 2023). cf-DNA can activate DNA sensors in immune cells, leading to increased expression of various inflammatory cytokines, such as TNF-α, IL-1β, and IL-6, exacerbating the progression of sepsis-associated acute lung injury (ALI) (Liu et al. 2021). Various clinical studies have shown that elevated levels of plasma cf-DNA, particularly cf-mtDNA, are associated with increased deterioration of sepsis-related lung injury. Targeting pulmonary cf-mtDNA may be an effective strategy for treating sepsis-associated ALI, as effective clearance of cf-mtDNA by deoxyribonuclease I (DNase-I) significantly improves survival and reduces lung injury (Qiao et al. 2022). Research has indicated that cfDNA is a promising candidate biomarker for the early diagnosis and monitoring of urinary tract infection (UTI) and is correlated with total extracellular DNA (ecDNA) levels, C-reactive protein (CRP), and procalcitonin in plasma and urine, enabling combined monitoring of the development of urosepsis (Mihaľová et al. 2023). cfDNA is highly accurate in the microbiological diagnosis of infectious endocarditis (IE) and may help determine the duration of individualized antibiotic therapy, allowing clinicians to estimate when a patient's infection has been adequately treated and potentially reducing the duration of antibiotic use (Eichenberger et al. 2023). Studies have confirmed that cfDNA levels are elevated not only in patients with sepsis compared to those in nonseptic controls or systemic inflammatory response syndrome (SIRS) ICU patients but also in nonsurvivors of sepsis compared to sepsis survivors. The data indicate that the sensitivity and specificity of cfDNA for diagnosing sepsis were 0.81 (95% CI 0.73–0.88, I2 = 8.9%) and 0.92 (95% CI 0.85–0.97, I2 = 0%), respectively. Measuring cfDNA levels in the early stages of ICU admission or upon hospitalization also predicts mortality, with a pooled AUC of 0.76 (95% CI 0.64–0.87) (Charoensappakit et al. 2023). Jing et al. (2022) also reported that cfDNA performs well in the early identification of septic patients and in the prediction of ICU patient prognosis, with AUCs of 0.992 (95% CI 0.969–1.000) and 0.802 (95% CI 0.605–0.999), respectively. Experimental methods for detecting cfDNA include polymerase chain reaction (PCR), a molecular diagnostic method, that is widely used for pathogen identification in sepsis patients. However, due to its low sensitivity, incomplete resistance information, and unclear assumptions about pathogens, it cannot meet the demands of complex clinical samples with unknown etiology (Ungerer et al. 2020). Metagenomic next-generation sequencing (mNGS) technology has been applied to diagnose infectious diseases by detecting microbial nucleic acids, but the results are complex and uncertain (Grumaz et al. 2016). Currently, the use of metagenomic next-generation sequencing (mNGS) through cfDNA sequencing for sepsis detection has a sensitivity of 68.1%, approximately 28% higher than that of blood culture, significantly improving the accuracy of sepsis patient diagnosis but at a high cost (Charoensappakit et al. 2023). As an emerging sepsis biomarker, cfDNA is stable in plasma and has been extensively studied for its ability to predict sepsis and monitor disease, although large-scale studies are still needed.

Adrenomedullin (ADM)

In sepsis, both vascular tone and vascular integrity are compromised, leading to the occurrence of shock. Adrenomedullin (ADM) is a vasoactive hormone that exerts vasodilatory effects by binding to receptors on endothelial and smooth muscle cells. ADM can regulate endothelial barrier function and stabilize the vascular endothelium (Geven et al. 2018) and there are two main methods for measuring ADM in peripheral blood. One is based on the mid-regional proadrenomedullin (MR-proADM) (Bergmann et al. 2005), and the other is the direct measurement of bioactive adrenomedullin (bio-ADM) (Marino et al. 2014). ADM is widely expressed in many organs and tissues. Plasma ADM concentrations are low in healthy individuals but significantly increase during pathological events, with changes in plasma concentrations proportional to the severity of the disease (Saeed et al. 2021), and numerous studies have reported an association between elevated ADM levels and poor prognosis in sepsis and septic shock patients (van Lier et al. 2020). However, ADM is rapidly cleared from circulation, making it difficult to detect. Therefore, the more stable MR-proADM directly reflects ADM levels and is used as asubstitute (Gille et al. 2017), and MR-proADM demonstrates high accuracy in the diagnosis of adult sepsis (AUC≥0.90) (Liang et al. 2023), with reports suggesting that MR-proADM levels are unaffected by pathogen type but rather reflect the degree of organ failure and disease severity, providing more extensive and reliable diagnostic and prognostic information than CRP and PCT (Valenzuela-Sánchez et al. 2016). Christ-Crain et al. (2005) also reported pro-ADM as a predictive marker for sepsis, with pro-ADM gradually increasing as critically ill patients progress from noninfection to sepsis, severe sepsis, and septic shock. Using 3.9 µg/L as a cutoff value, the sensitivity and specificity of diagnosing sepsis with pro-ADM were 83.3% and 87.8%, respectively, demonstrating superior diagnostic accuracy to CRP and PCT. Additionally, pro-ADM levels in the sepsis mortality group were significantly greater than those in the survival group, with a receiver operating characteristic (ROC) area under the curve (AUC) of 0.81 for predicting adverse outcomes in sepsis patients. Bio-ADM can be directly detected and exhibits strong prognostic monitoring capabilities in sepsis patients in the emergency department, with increased bio-ADM levels associated with mortality, an increased organ failure count, an increased ICU admission rate, and a decreased emergency department discharge rate (Lundberg et al. 2022). Given that ADM is a key hormone involved in regulating vascular tone and endothelial barrier function, increased bio-ADM within blood vessels may have therapeutic value in sepsis. This has led to studies of the nonneutralizing anti-ADM antibody Adrecizumab in humans (Karakas et al. 2020), which has been shown to improve outcomes in preclinical sepsis models, but further research is needed.

Systemic immune-inflammation index

Research has shown that inflammatory markers derived from blood cell counts, including the neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR), are more sensitive inflammatory biomarkers than individual blood cell counts are. The neutrophil–lymphocyte ratio (NLR) and platelet–lymphocyte ratio (PLR) can be obtained through routine blood tests, which are cost-effective and more accessible, and can be used for the early identification of sepsis. A meta-analysis showed that a higher NLR is associated with poor prognosis in septic patients (mean hazard ratio (HR) 1.75), with nonsurvivors having a greater NLR than survivors of sepsis (Huang et al. 2020). In a single-center prospective observational study of intensive care unit (ICU) septic patients, the NLR (9.53±2.31) was correlated with the severity of sepsis according to the SOFA score (R = 0.65), with a significantly greater NLR in septic shock patients (10.31±2.32), particularly when the NLR exceeded 10,indicating the potential value of the NLR in assessing the severity of sepsis (Drăgoescu et al. 2021). The NLR is also an independent risk factor for 28-day mortality in bloodstream sepsis patients, with its elevation related to disease severity, and ROC curve analysis revealed that both PCT and the NLR had an AUC > 0.7 in predicting the risk of patient death within 28 days (Liang and Yu 2022). However, trauma, surgery, pancreatitis, and rheumatic diseases can cause nonspecific increases in the NLR (Karakonstantis et al. 2017). A study using data from the Beth Israel Deaconess Medical Center database at Harvard University in the United States showed that the PLR is a factor affecting the 28-day mortality of ICU septic patients (Shen et al. 2019). High PLR values indicate more severe sepsis and poorer prognosis in septic patients, with one study showing that when the PLR exceeds 112, the sensitivity for predicting ICU mortality is 83%, with a specificity of 52% (AUC:0.76, p<0.001) (Karagoz and Yoldas 2019). Therefore, high NLR and PLR values indicate severe illness and poor prognosis in septic patients. However, due to their limited predictive value, these biomarkers cannot accurately determine the severity of infection. Therefore, prognostic indicators based on neutrophil, lymphocyte, and platelet counts are expected to be more reliable than markers based on single factors (Lippi et al. 2020). Immune cells play crucial roles in the development of inflammation. The systemic immune-inflammation index (SII) is a novel inflammatory biomarker derived from peripheral blood cell counts, and is calculated as the platelet (P) count multiplied by the neutrophil (N) count divided by the lymphocyte (L) count (SII = P × N/L) (Hu et al. 2014). The systemic immune-inflammation index (SII), first proposed by Hu in 2014 (Hu et al. 2014), is a tool used to assess the prognosis of hepatocellular carcinoma (HCC) patients by combining neutrophil, lymphocyte, and platelet counts. The SII includes three types of inflammatory cells, neutrophils, lymphocytes, and platelets, and provides an objective and intuitive reflection of the relationship between inflammation and the host immune response. A higher SII indicates higher platelet/neutrophil levels and/or lower lymphocyte counts. An elevated SII indirectly signifies a weakened immune response and enhanced inflammation, which are associated with a poorer prognosis. High SII values correlate with other inflammatory markers, such as CRP, ferritin, procalcitonin, and cytokines(Muhammad et al. 2021). The systemic immune-inflammation index (SII), which is derived from peripheral blood cell counts, has been widely used to predict and evaluate various cancer prognostic indicators (Qiu et al. 2021). It has also been applied in neurological diseases, where an elevated SII on the first day can be used to assess the prognosis of brain hemorrhage (Li et al. 2021). Abnormal blood lipid levels are often linked to inflammation, and research has shown a significant positive correlation between the SII and hyperlipidemia (Mahemuti et al. 2023). The SII can predict the risk of cardiovascular diseases and assess the prognosis of cardiovascular diseases (Ye et al. 2022). Recently, the SII has shown increased utility as a marker of systemic inflammation. In acute pancreatitis, the interplay between neutrophils and reactive oxygen species (ROS) promotes the progression of pancreatitis, and platelets play a direct role in the systemic inflammatory process. The SII can serve as a potential indicator for predicting the severity of acute pancreatitis (Liu et al. 2021). The SII has also been found to be effective at predicting the prognosis of patients with odontogenic septicaemia (Pricop et al. 2022). Zhongwei Huang's team conducted a retrospective cohort study based on the MIMIC-IV database and found that in critically ill patients with sepsis, there was a J-shaped relationship between the SII and short-term mortality. Both low and high SII values are associated with an increased risk of short-term mortality, with the lowest 28-day mortality risk associated with a SII level of 774.46*10⁹/L (Jiang et al. 2023).

Monocyte distribution width (MDW)

The monocyte distribution width (MDW) has recently emerged as a promising biomarker for sepsis, especially in acute care settings such as emergency departments and intensive care units. Early data on the potential importance of the MDW were gathered, indicating that when activated in bacteremia patients, monocytes enlarge, and this infection-related size change can be easily observed by monitoring the distribution of monocytes in coulter chambers (Kou et al. 2020). Crouser et al. (2017) reported for the first time that in emergency department septic patients, the MDW can be used to measure monocyte activity and morphological changes in early inflammatory responses, indicating that the MDW enhances the initial detection of sepsis. Karam et al. (Aktas et al. 2023). reported that the overall ROC area of MDW (0.790) was greater than that of PCT (0.760), indicating that the diagnostic accuracy of MDW is greater than that of PCT, and the MDW biomarker with a cutoff value > 20 can be used as a diagnostic marker for septic patients. Luisa et al. (Agnello et al. 2022). reported that the overall sensitivity and specificity of the MDW for diagnosing sepsis patients were 0.838 and 0.704,respectively (Monocyte distribution width (MDW) as a screening tool for the early detection of sepsis: systematic review and meta-analysis). Introducing MDW into clinical practice is very attractive because its assessment is rapid, simple, and cost-effective,and does not require additional blood samples, similar to measuring other biomarkers. It can be used in conjunction with the SII for the early identification and monitoring of sepsis patients throughout the course of sepsis.

Conclusion

Sepsis is a complex process involving various changes in cellular and cytokine levels, making early diagnosis crucial for reducing mortality and improving prognosis. The identification of key cytokines involved in the occurrence and progression of septicaemia is essential. Many novel biomarkers for septicemia have been discovered, and there is potential for finding even more accurate and sensitive biomarkers in the future, which could include specific proteins, nucleic acid sequences, metabolites, or cytokines. Additionally, with the rapid advancement of science and technology, the application of artificial intelligence and machine learning in biomarker detection techniques is continually improving. This may lead to the development of faster, more sensitive, convenient, and cost-effective detection methods, facilitating better utilization of biomarkers in clinical practice. This advancement could enable earlier identification of septic patients, guide patient treatment, and improve patient prognosis. However, whether traditional or novel, there is yet to be a single biomarker with sufficiently high sensitivity and specificity for the diagnosis, monitoring, and treatment of sepsis. Therefore, the combination of multiple biomarkers is necessary to enhance the sensitivity and specificity of early sepsis diagnosis. Additionally, there is still a need to find an ideal monitoring model to guide treatment, improve sepsis prognosis, and reduce mortality.

Statement of author contribution

Analysis and interpretation of the data: Hong Zhang, Junpeng Zhao, Mengxiang Huang Mingbing Xiao

Drafting of the paper: Hong Zhang Junpeng Zhao

Critically revising the manuscript for intellectual content: Zhiping Wang, Zhonghua Tan, Feng Jiang, Yuhong Zhou Mingbing Xiao, Linhua Wang

The final approval of the version to be published: Mingbing Xiao

All the authors agree to be accountable for all aspects of the work.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

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

Additional information

Funding

This study was supported by grants from the National Natural Science Foundation of China (No. 82272624), the Natural Science Foundation of Jiangsu Province (No. BK20211105), Jiangsu Provincial Research Hospital (YJXYY202204), the Plan of Jiangsu Provincial Medical Key Discipline (No. ZDXK202240), the Social Development Foundation of Nantong City (MS22022044, JC22022001), the Health Project of Nantong City (MS2023009), the Jiangsu Administration of Traditional Chinese Medicine (JD2022SZ07), and the Jiangsu Elderly Health Research Project (LX2021012).