Application of Artificial Intelligence to Advance Individualized Diagnosis and Treatment in Emergency and Critical Care Medicine.

Emergency and critical illnesses refer to severe diseases or conditions characterized by rapid changes in health that may endanger life within a short period [...].

Genetic correlations and causal relationships between cardio-metabolic traits and sepsis.

Cardio-metabolic traits have been reported to be associated with the development of sepsis. It is, however, unclear whether these co-morbidities reflect causal associations, shared genetic heritability, or are confounded by environmental factors. We performed three analyses to explore the relationships between cardio-metabolic traits and sepsis. Mendelian randomization (MR) study to evaluate the causal effects of multiple cardio-metabolic traits on sepsis. Global genetic correlation analysis to explore the correlations between cardio-metabolic traits and sepsis. Local genetic correlation (GC) analysis to explore shared genetic heritability between cardio-metabolic traits and sepsis. Some loci were further examined for related genes responsible for the causal relationships. Genetic associations were obtained from the UK Biobank data or published large-scale genome-wide association studies with sample sizes between 200,000 to 750,000. In MR, we found causality between BMI and sepsis (OR: 1.53 [1.4-1.67]; p < 0.001). Body mass index (BMI), which is confirmed by sensitivity analyses and multivariable MR adjusting for confounding factors. Global GC analysis showed a significant correlation between BMI and sepsis (rg = 0.55, p < 0.001). More cardio-metabolic traits were identified to be correlated to the sepsis onset such as CRP (rg = 0.37, p = 0.035), type 2 diabetes (rg = 0.33, p < 0.001), HDL (rg = - 0.41, p < 0.001), and coronary artery disease (rg = 0.43, p < 0.001). Local GC revealed some shared genetic loci responsible for the causality. The top locus 1126 was located at chromosome 7 and comprised genes HIBADH, JAZF1, and CREB5. The present study provides evidence for an independent causal effect of BMI on sepsis. Further detailed analysis of the shared genetic heritability between cardio-metabolic traits and sepsis provides the opportunity to improve the preventive strategies for sepsis.

Genomic surveillance indicates clonal replacement of hypervirulent Klebsiella pneumoniae ST881 and ST29 lineage strains in vivo.

The emergence of hypervirulent Klebsiella pneumoniae (hvKp) poses a significant public health threat, particularly regarding its carriage in the healthy population. However, the genomic epidemiological characteristics and population dynamics of hvKp within a single patient across distinct infection episodes remain largely unknown. This study aimed to investigate the clonal replacement of hvKp K2-ST881 and K54-ST29 lineage strains in a single patient experiencing multiple-site infections during two independent episodes. Two strains, designated EDhvKp-1 and EDhvKp-2, were obtained from blood and cerebrospinal fluid during the first admission, and the strain isolated from blood on the second admission was named EDhvKp-3. Whole-genome sequencing, utilizing both short-read Illumina and long-read Oxford Nanopore platforms, was conducted. In silico multilocus sequence typing (MLST), identification of antimicrobial resistance and virulence genes, and the phylogenetic relationship between our strains and other K. pneumoniae ST881 and ST29 genomes retrieved from the public database were performed. Virulence potentials were assessed through a mouse lethality assay. Our study indicated that the strains were highly susceptible to multiple antimicrobial agents. Plasmid sequence analysis confirmed that both virulence plasmids, pEDhvKp-1 (166,008 bp) and pEDhvKp-3 (210,948 bp), belonged to IncFIB type. Multiple virulence genes, including rmpA, rmpA2, rmpC, rmpD, iroBCDN, iucABCD, and iutA, were identified. EDhvKp-1 and EDhvKp-2 showed the closest relationship to strain 502 (differing by 51 SNPs), while EDhvKp-3 exhibited 69 SNPs differences compared to strain TAKPN-1, which all recovered from Chinese patients in 2020. In the mouse infection experiment, both ST881 EDhvKp-1 and ST29 EDhvKp-3 displayed similar virulence traits, causing 90 and 100% of the mice to die within 72 h after intraperitoneal infection, respectively. Our study expands the spectrum of hvKp lineages and highlights genomic alterations associated with clonal switching between two distinct lineages of hvKP that successively replaced each other in vivo. The development of novel strategies for the surveillance, diagnosis, and treatment of high-risk hvKp is urgently needed.

Exploring disease axes as an alternative to distinct clusters for characterizing sepsis heterogeneity.

PURPOSE: Various studies have analyzed sepsis subtypes, yet the reproducibility of such results remains unclear. This study aimed to determine the reproducibility of sepsis subtypes across multiple cohorts. METHODS: The study examined 63,547 sepsis patients from six distinct cohorts who had similar sepsis-related characteristics (vital signs, lactate, sequential organ failure assessment score, bilirubin, serum, urine output, and Glasgow coma scale). Identical cluster analysis techniques were used, employing 27 clustering schemes, and normalized mutual information (NMI), a metric ranging from 0 to 1 with higher values indicating better concordance, was employed to quantify the clustering solutions' reproducibility. Principal component analysis (PCA) was utilized to obtain the disease axis, and its uniformity across cohorts was evaluated through patterns of feature loading and correlation. RESULTS: The reproducibility of sepsis clustering subtypes across the various studies was modest (median NMI ranging from 0.08 to 0.54). The top-down transfer learning method (model trained on cohorts with greater severity was transferred to cohorts with lower severity score) had a higher NMI value than the bottom-up approach (median [Q1, Q3]: 0.64 [0.49, 0.78] vs. 0.23 [0.2, 0.31], p < 0.001). The reproducibility was greater when the transfer solution was performed within United States (US) cohorts. The PCA analysis revealed that the correlation pattern between variables was consistent across all cohorts, and the first two disease axes were the "shock axis" and "systemic inflammatory response syndrome (SIRS) axis." CONCLUSIONS: Cluster analysis of sepsis patients across various cohorts showed modest reproducibility. Sepsis heterogeneity is better characterized through continuous disease axes that coexist to varying degrees within the same individual instead of mutually exclusive subtypes.

Insight into the history and trends of surgical simulation training in education: a bibliometric analysis.

BACKGROUND: Surgical simulation training enables surgeons to acquire clinical experience or skills from the operating room to the simulation environment. Historically, it has changed with advances in science and technology. Moreover, no previous study has analyzed this field from the bibliometric analysis dimension. The study aimed to review changes in surgical simulation training worldwide using bibliometric software. MATERIALS AND METHODS: Two searches were performed on the core collection database, Web of Science, regarding data from 1991 to the end of 2020 using three topic words (surgery, training, and simulation). From 1 January 2000, to 15 May 2022, the keyword 'robotic' was added for the hotspot exploration. The data were chiefly analyzed by publication date, country, author(s), and keywords using bibliometric software. RESULTS: A total of 5285 articles were initially analyzed, from which it was clear that laparoscopic skill, three-dimensional printing, and virtual reality were the main focuses during those study periods. Subsequently, 348 publications on robotic surgery training were identified. CONCLUSION: This study systematically summarizes the current status in the field of surgical simulation training and provides insights into the research focuses and future hotspot in a global context.

Corrigendum: Adiponectin modified BMSCs alleviate heart fibrosis via inhibition TGF-beta1/smad in diabetic rats.

[This corrects the article DOI: 10.3389/fcell.2021.644160.].

The application of artificial intelligence in the management of sepsis.

Sepsis is a complex and heterogeneous syndrome that remains a serious challenge to healthcare worldwide. Patients afflicted by severe sepsis or septic shock are customarily placed under intensive care unit (ICU) supervision, where a multitude of apparatus is poised to produce high-granularity data. This reservoir of high-quality data forms the cornerstone for the integration of AI into clinical practice. However, existing reviews currently lack the inclusion of the latest advancements. This review examines the evolving integration of artificial intelligence (AI) in sepsis management. Applications of artificial intelligence include early detection, subtyping analysis, precise treatment and prognosis assessment. AI-driven early warning systems provide enhanced recognition and intervention capabilities, while profiling analyzes elucidate distinct sepsis manifestations for targeted therapy. Precision medicine harnesses the potential of artificial intelligence for pathogen identification, antibiotic selection, and fluid optimization. In conclusion, the seamless amalgamation of artificial intelligence into the domain of sepsis management heralds a transformative shift, ushering in novel prospects to elevate diagnostic precision, therapeutic efficacy, and prognostic acumen. As AI technologies develop, their impact on shaping the future of sepsis care warrants ongoing research and thoughtful implementation.

Gene signature for the prediction of the trajectories of sepsis-induced acute kidney injury.

BACKGROUND: Acute kidney injury (AKI) is a common complication in sepsis. However, the trajectories of sepsis-induced AKI and their transcriptional profiles are not well characterized. METHODS: Sepsis patients admitted to centres participating in Chinese Multi-omics Advances In Sepsis (CMAISE) from November 2020 to December 2021 were enrolled, and gene expression in peripheral blood mononuclear cells was measured on Day 1. The renal function trajectory was measured by the renal component of the SOFA score (SOFArenal) on Days 1 and 3. Transcriptional profiles on Day 1 were compared between these renal function trajectories, and a support vector machine (SVM) was developed to distinguish transient from persistent AKI. RESULTS: A total of 172 sepsis patients were enrolled during the study period. The renal function trajectory was classified into four types: non-AKI (SOFArenal = 0 on Days 1 and 3, n = 50), persistent AKI (SOFArenal > 0 on Days 1 and 3, n = 62), transient AKI (SOFArenal > 0 on Day 1 and SOFArenal = 0 on Day 3, n = 50) and worsening AKI (SOFArenal = 0 on Days 1 and SOFArenal > 0 on Day 3, n = 10). The persistent AKI group showed severe organ dysfunction and prolonged requirements for organ support. The worsening AKI group showed the least organ dysfunction on day 1 but had higher serum lactate and prolonged use of vasopressors than the non-AKI and transient AKI groups. There were 2091 upregulated and 1,902 downregulated genes (adjusted p < 0.05) between the persistent and transient AKI groups, with enrichment in the plasma membrane complex, receptor complex, and T-cell receptor complex. A 43-gene SVM model was developed using the genetic algorithm, which showed significantly greater performance predicting persistent AKI than the model based on clinical variables in a holdout subset (AUC: 0.948 [0.912, 0.984] vs. 0.739 [0.648, 0.830]; p < 0.01 for Delong's test). CONCLUSIONS: Our study identified four subtypes of sepsis-induced AKI based on kidney injury trajectories. The landscape of host response aberrations across these subtypes was characterized. An SVM model based on a gene signature was developed to predict renal function trajectories, and showed better performance than the clinical variable-based model. Future studies are warranted to validate the gene model in distinguishing persistent from transient AKI.

Single-cell transcriptome profiling reveals heterogeneous neutrophils with prognostic values in sepsis.

Neutrophils constitute the largest proportion of nucleated peripheral blood cells, and neutrophils have substantial heterogeneity. We profiled nearly 300,000 human peripheral blood cells in this study using single-cell RNA sequencing. A large proportion (>50%) of these cells were annotated as neutrophils. Neutrophils were further clustered into four subtypes, including Neu1, Neu2, Neu3, and Neu4. Neu1 is characterized by high expression of MMP9, HP, and RGL4. Neu1 was associated with septic shock and significantly correlated with the sequential organ failure assessment (SOFA) score. A gene expression module in Neu1 named Neu1_C (characterized by expression of NFKBIA, CXCL8, G0S2, and FTH1) was highly predictive of septic shock with an area under the curve of 0.81. The results were extensively validated in external bulk datasets by using single-cell deconvolution methods. In summary, our study establishes a general framework for studying neutrophil-related mechanisms, prognostic biomarkers, and potential therapeutic targets for septic shock.

Identifying Novel Clusters of Patients With Prolonged Mechanical Ventilation Using Trajectories of Rapid Shallow Breathing Index.

Objective: Patients with prolonged mechanical ventilation (PMV) are comprised of a heterogeneous population, creating great challenges for clinical management and study design. The study aimed to identify subclusters of PMV patients based on trajectories of rapid shallow breathing index (RSBI), and to develop a machine learning model to predict the cluster membership based on baseline variables. Methods: This was a retrospective cohort study conducted in respiratory care center (RCC) at a tertiary academic medical center. The RCC referral criteria were patients with mechanical ventilation for at least 21 days with stable hemodynamic and oxygenation status. Patients admitted to the RCC from April 2009 to December 2020 were screened. Two-step clustering through linear regression modeling and k-means was employed to find clusters of the trajectories of RSBI. The number of clusters was chosen by statistical metrics and domain expertise. A gradient boosting machine (GBM) was trained, exploiting variables on RCC admission, to predict cluster membership. Results: A total of 1371 subjects were included in the study. Four clusters were identified: cluster A showed persistently high RSBI; cluster B was characterized by a constant low RSBI over time; Cluster C was characterized by increasing RSBI; and cluster D showed a declining RSBI. Cluster A showed the highest mortality rate (72%), followed by cluster D (63%), C (62%) and B (61%; p = 0.005 for comparison between 4 clusters). GBM was able to predict cluster membership with an accuracy of > 0.95 in ten-fold cross validation. Highly ranked variables for the prediction of clusters included thyroid-stimulating hormone (TSH), cortisol, platelet, free thyroxine (T4) and serum magnesium. Conclusions: Patients with PMV are composed of a heterogeneous population that can be classified into four clusters by using trajectories of RSBI. These clusters can be easily predicted with baseline clinical variables.

Effectiveness of automated alerting system compared to usual care for the management of sepsis.

There is a large body of evidence showing that delayed initiation of sepsis bundle is associated with adverse clinical outcomes in patients with sepsis. However, it is controversial whether electronic automated alerts can help improve clinical outcomes of sepsis. Electronic databases are searched from inception to December 2021 for comparative effectiveness studies comparing automated alerts versus usual care for the management of sepsis. A total of 36 studies are eligible for analysis, including 6 randomized controlled trials and 30 non-randomized studies. There is significant heterogeneity in these studies concerning the study setting, design, and alerting methods. The Bayesian meta-analysis by using pooled effects of non-randomized studies as priors shows a beneficial effect of the alerting system (relative risk [RR]: 0.71; 95% credible interval: 0.62 to 0.81) in reducing mortality. The automated alerting system shows less beneficial effects in the intensive care unit (RR: 0.90; 95% CI: 0.73-1.11) than that in the emergency department (RR: 0.68; 95% CI: 0.51-0.90) and ward (RR: 0.71; 95% CI: 0.61-0.82). Furthermore, machine learning-based prediction methods can reduce mortality by a larger magnitude (RR: 0.56; 95% CI: 0.39-0.80) than rule-based methods (RR: 0.73; 95% CI: 0.63-0.85). The study shows a statistically significant beneficial effect of using the automated alerting system in the management of sepsis. Interestingly, machine learning monitoring systems coupled with better early interventions show promise, especially for patients outside of the intensive care unit.

The landscape of isoform switches in sepsis: a multicenter cohort study.

Sepsis is caused by an uncontrolled inflammatory response, whose underlying mechanisms are not fully understood. It is well known that the majority of human genes can be expressed as alternative isoforms. While isoform switching is implicated in many diseases and is particularly prominent in cancer, it has never been reported in the context of sepsis. Patients presented to the emergency department of three tertiary care hospitals from January 2020 to December 2020 were enrolled. Clinical variables and genome-wide transcriptome of peripheral blood mononuclear cells (PBMC) were obtained. Isoform switching analysis were performed to identify significant isoform switches and relevant biological consequences. A total of 48 subjects with sepsis, involving 42 survivors and 6 non-survivors, admitted to the emergency department of three tertiary care hospitals were enrolled in this study. PBMCs were extracted for RNA sequencing (RNA-seq). Patients (n = 4) with mild stroke or acute coronary syndrome without infection were enrolled in this study as controls. The most frequent functional changes resulting from isoform switching were changes affecting the open reading frame, protein domains and intron retention. Many genes without differences in gene expression showed significant isoform switching. Many genes with significant isoform switches ([Formula: see text]> 0.1) were associated with higher mortality risk, including PIGS, CASP3, LITAF, HBB and RUVBL2. The study for the first time described the landscape of isoform switching in sepsis, including differentially expressed isoform fractions between patients with and without sepsis and survivors and nonsurvivors. The biological consequences of isoform switching, including protein domain loss, signal peptide gain, and intron retention, were identified.

Integrated Single Cell and Bulk RNA-Seq Analysis Revealed Immunomodulatory Effects of Ulinastatin in Sepsis: A Multicenter Cohort Study.

Sepsis is a leading cause of morbidity and mortality in the intensive care unit, which is caused by unregulated inflammatory response leading to organ injuries. Ulinastatin (UTI), an immunomodulatory agent, is widely used in clinical practice and is associated with improved outcomes in sepsis. But its underlying mechanisms are largely unknown. Our study integrated bulk and single cell RNA-seq data to systematically explore the potential mechanisms of the effects of UTI in sepsis. After adjusting for potential confounders in the negative binomial regression model, there were more genes being downregulated than being upregulated in the UTI group. These down-regulated genes were enriched in the neutrophil involved immunity such as neutrophil activation and degranulation, indicating the immunomodulatory effects of UTI is mediated via regulation of neutrophil activity. By deconvoluting the bulk RNA-seq samples to obtain fractions of cell types, the Myeloid-derived suppressor cells (MDSC) were significantly expanded in the UTI treated samples. Further cell-cell communication analysis revealed some signaling pathways such as ANEEXIN, GRN and RESISTIN that might be involved in the immunomodulatory effects of UTI. The study provides a comprehensive reference map of transcriptional states of sepsis treated with UTI, as well as a general framework for studying UTI-related mechanisms.

Effectiveness of anisodamine for the treatment of critically ill patients with septic shock: a multicentre randomized controlled trial.

BACKGROUND: Septic shock is characterized by an uncontrolled inflammatory response and microcirculatory dysfunction. There is currently no specific agent for treating septic shock. Anisodamine is an agent extracted from traditional Chinese medicine with potent anti-inflammatory effects. However, its clinical effectiveness remains largely unknown. METHODS: In a multicentre, open-label trial, we randomly assigned adults with septic shock to receive either usual care or anisodamine (0.1-0.5 mg per kilogram of body weight per hour), with the anisodamine doses adjusted by clinicians in accordance with the patients' shock status. The primary end point was death on hospital discharge. The secondary end points were ventilator-free days at 28 days, vasopressor-free days at 28 days, serum lactate and sequential organ failure assessment (SOFA) score from days 0 to 6. The differences in the primary and secondary outcomes were compared between the treatment and usual care groups with the χ2 test, Student's t test or rank-sum test, as appropriate. The false discovery rate was controlled for multiple testing. RESULTS: Of the 469 patients screened, 355 were assigned to receive the trial drug and were included in the analyses-181 patients received anisodamine, and 174 were in the usual care group. We found no difference between the usual care and anisodamine groups in hospital mortality (36% vs. 30%; p = 0.348), or ventilator-free days (median [Q1, Q3], 24.4 [5.9, 28] vs. 26.0 [8.5, 28]; p = 0.411). The serum lactate levels were significantly lower in the treated group than in the usual care group after day 3. Patients in the treated group were less likely to receive vasopressors than those in the usual care group (OR [95% CI] 0.84 [0.50, 0.93] for day 5 and 0.66 [0.37, 0.95] for day 6). CONCLUSIONS: There is no evidence that anisodamine can reduce hospital mortality among critically ill adults with septic shock treated in the intensive care unit. Trial registration ClinicalTrials.gov ( NCT02442440 ; Registered on 13 April 2015).

Individualized Mechanical power-based ventilation strategy for acute respiratory failure formalized by finite mixture modeling and dynamic treatment regimen.

BACKGROUND: Mechanical ventilation (MV) is the key to the successful treatment of acute respiratory failure (ARF) in the intensive care unit (ICU). The study aims to formalize the concept of individualized MV strategy with finite mixture modeling (FMM) and dynamic treatment regime (DTR). METHODS: ARF patients requiring MV for over 48 h from 2008 to 2019 were included. FMM was conducted to identify classes of ARF. Static and dynamic mechanical power (MP_static and MP_dynamic) and relevant clinical variables were calculated/collected from hours 0 to 48 at an interval of 8 h. Δ M P was calculated as the difference between actual and optimal MP. FINDINGS: A total of 8768 patients were included for analysis with a mortality rate of 27%. FFM identified three classes of ARF, namely, the class 1 (baseline), class 2 (critical) and class 3 (refractory respiratory failure). The effect size of MP_static on mortality is the smallest in class 1 (HR for every 5 Joules/min increase: 1.29; 95% CI: 1.15 to 1.45; p < 0.001) and the largest in class 3 (HR for every 5 Joules/min increase: 1.83; 95% CI: 1.52 to 2.20; p < 0.001). INTERPRETATION: MP has differing therapeutic effects for subtypes of ARF. Optimal MP estimated by DTR model may help to improve survival outcome. FUNDING: The study was funded by Health Science and Technology Plan of Zhejiang Province (2021KY745), Key Research & Development project of Zhejiang Province (2021C03071) and Yilu "Gexin" - Fluid Therapy Research Fund Project (YLGX-ZZ-2,020,005).

Latent transition analysis of cardiac arrest patients treated in the intensive care unit.

BACKGROUND AND OBJECTIVE: Post-cardiac arrest (CA) syndrome is heterogenous in their clinical presentations and outcomes. This study aimed to explore the transition and stability of subphenotypes (profiles) of CA treated in the intensive care unit (ICU). PATIENTS AND METHODS: Clinical features of CA patients on day 1 and 3 after ICU admission were modeled by latent transition analysis (LTA) to explore the transition between subphenotypes over time. The association between different transition patterns and mortality outcome was explored using multivariable logistic regression. RESULTS: We identified 848 eligible patients from the database. The LPA identified three distinct subphenotypes: Profile 1 accounted for the largest proportion (73%) and was considered as the baseline subphenotype. Profile 2 (13%) was characterized by brain injury and profile 3 (14%) was characterized by multiple organ dysfunctions. The same three subphenotypes were identified on day 3. The LTA showed consistent subphenotypes. A majority of patients in profile 2 (72%) and 3 (82%) on day 1 switched to profile 1 on day 3. In the logistic regression model, patients in profile 1 on day 1 transitioned to profile 3 had worse survival outcome than those continue to remain in profile 1 (OR: 20.64; 95% CI: 6.01 to 70.94; p < 0.001) and transitioned to profile 2 (OR: 8.42; 95% CI: 2.22 to 31.97; p = 0.002) on day 3. CONCLUSION: The study identified three subphenotypes of CA, which was consistent on day 1 and 3 after ICU admission. Patients who transitioned to profile 3 on day 3 had significantly worse survival outcome than those remained in profile 1 or 2.

Adiponectin Modified BMSCs Alleviate Heart Fibrosis via Inhibition TGF-beta1/Smad in Diabetic Rats.

Background: Accumulating evidence suggested that bone marrow mesenchymal stem cells (BMSCs) have therapeutic potential for diabetes and heart diseases. However, the effects of BMSC on reducing myocardial fibrosis need to be optimized. This study aimed to investigate the mechanism of adiponectin (APN) modified BMSCs on myocardial fibrosis in diabetic model in vivo and in vitro. Methods: The high-fat diet combined with streptozotocin (STZ) injection were used to induced diabetic rat model. H9c2 cells were cultured under a high glucose medium as in vitro model. The BMSCs were modified by APN plasmid or APN small interfering RNA (siRNA), then transplanted to the diabetic rats by a single tail-vein injection, or co-cultured with H9c2 cells. Results: We demonstrated that diabetic rats showed typical diabetic symptoms, such as decreased cardiac function, accumulation of pathological lesions and collagen expression. However, these impairments were significantly prevented by the APN modified BMSCs treatment while no effects on APN siRNA modified BMSCs treated diabetic rats. Moreover, we confirmed that APN modified BMSCs could attenuate the expression of TGF-beta1/smad to suppress the myocardial fibrosis in the diabetic rats and high glucose induced H9c2 cells. Conclusion: The present results for the first time showed that APN modified BMSCs exerted protection on cardiac fibrosis via inhibiting TGF-beta1/smad signal pathway in diabetic rats. Our findings suggested that APN modified BMSCs might be a novel and optimal therapy for the diabetic cardiomyopathy in future.

Deep learning-based clustering robustly identified two classes of sepsis with both prognostic and predictive values.

BACKGROUND: Sepsis is a heterogenous syndrome and individualized management strategy is the key to successful treatment. Genome wide expression profiling has been utilized for identifying subclasses of sepsis, but the clinical utility of these subclasses was limited because of the classification instability, and the lack of a robust class prediction model with extensive external validation. The study aimed to develop a parsimonious class model for the prediction of class membership and validate the model for its prognostic and predictive capability in external datasets. METHODS: The Gene Expression Omnibus (GEO) and ArrayExpress databases were searched from inception to April 2020. Datasets containing whole blood gene expression profiling in adult sepsis patients were included. Autoencoder was used to extract representative features for k-means clustering. Genetic algorithms (GA) were employed to derive a parsimonious 5-gene class prediction model. The class model was then applied to external datasets (n = 780) to evaluate its prognostic and predictive performance. FINDINGS: A total of 12 datasets involving 1613 patients were included. Two classes were identified in the discovery cohort (n = 685). Class 1 was characterized by immunosuppression with higher mortality than class 2 (21.8% [70/321] vs. 12.1% [44/364]; p < 0.01 for Chi-square test). A 5-gene class model (C14orf159, AKNA, PILRA, STOM and USP4) was developed with GA. In external validation cohorts, the 5-gene class model (AUC: 0.707; 95% CI: 0.664 - 0.750) performed better in predicting mortality than sepsis response signature (SRS) endotypes (AUC: 0.610; 95% CI: 0.521 - 0.700), and performed equivalently to the APACHE II score (AUC: 0.681; 95% CI: 0.595 - 0.767). In the dataset E-MTAB-7581, the use of hydrocortisone was associated with increased risk of mortality (OR: 3.15 [1.13, 8.82]; p = 0.029) in class 2. The effect was not statistically significant in class 1 (OR: 1.88 [0.70, 5.09]; p = 0.211). INTERPRETATION: Our study identified two classes of sepsis that showed different mortality rates and responses to hydrocortisone therapy. Class 1 was characterized by immunosuppression with higher mortality rate than class 2. We further developed a 5-gene class model to predict class membership. FUNDING: The study was funded by the National Natural Science Foundation of China (Grant No. 81,901,929).