Derivation, Validation, and Potential Treatment Implications of Novel Clinical Phenotypes for Sepsis

Redefining clinical venous thromboembolism phenotypes: a novel approach using latent class analysis

BACKGROUND

Patients with venous thromboembolism (VTE) are commonly classified by the presence or absence of provoking factors at the time of VTE to guide treatment decisions. This approach may not capture the heterogeneity of the disease and its prognosis.

OBJECTIVES

To evaluate clinically important novel phenotypic clusters among patients with VTE without cancer and to explore their association with anticoagulant treatment and clinical outcomes.

METHODS

Latent class analysis was performed with 18 baseline clinical variables in 3062 adult patients with VTE without active cancer participating in PREFER in VTE, a noninterventional disease registry. The derived latent classes were externally validated in a post hoc analysis of Hokusai-VTE (n = 6593), a randomized trial comparing edoxaban with warfarin. The associations between cluster membership and anticoagulant treatment, recurrent VTE, bleeding, and mortality after initial treatment were studied.

RESULTS

The following 5 clusters were identified: young men cluster (n = 1126, 37%), young women cluster (n = 215, 7%), older people cluster (n = 1106, 36%), comorbidity cluster (n = 447, 15%), and history of venous thromboembolism cluster (n = 168, 5%). Patient characteristics varied by age, sex, medical history, and treatment patterns. Consistent clusters were evident on external validation. In Cox proportional hazard models, recurrence risk was lower in the young women cluster (hazard ratio [HR], 0.27; 95% CI, 0.12-0.61) compared with the comorbidity cluster, after adjusting for extended anticoagulation. The risk of bleeding was lower in young men, young women, and older people clusters (HR, 0.50; 95% CI, 0.38-0.66; HR, 0.23; 95% CI, 0.11-0.46; and HR, 0.55; 95% CI 0.41-0.73, respectively).

CONCLUSION

The heterogeneity of VTE cases extends beyond the distinction between provoked and unprovoked VTE.

Characterization of Mortality by Sepsis Source in Patients Admitted to the Surgical Intensive Care Unit

INTRODUCTION

The impact of infectious source on sepsis outcomes for surgical patients is unclear. The objective of this study was to evaluate the association between sepsis sources and cumulative 90-d mortality in patients admitted to the surgical intensive care unit (SICU) with sepsis.

METHODS

All patients admitted to the SICU at an academic institution who met sepsis criteria (2014-2019, n = 1296) were retrospectively reviewed. Classification of source was accomplished through a chart review and included respiratory (RT, n = 144), intra-abdominal (IA, n = 859), skin and soft tissue (SST, n = 215), and urologic (UR, n = 78). Demographics, comorbidities, and clinical presentation were compared. Outcomes included 90-d mortality, respiratory and renal failure, length of stay, and discharge disposition. Cox-proportional regression was used to model predictors of mortality; P < 0.05 was significant.

RESULTS

Patients with SST were younger, more likely to be diabetic and obese, but had the lowest total comorbidities. Median admission sequential organ failure assessment scores were highest for IA and STT and lowest in urologic infections. Cumulative 90-d mortality was highest for IA and RT (35% and 33%, respectively) and lowest for SST (20%) and UR (8%) (P < 0.005). Compared to the other categories, UR infections had the lowest SICU length of stay and the highest discharge-to-home (57%, P < 0.0005). Urologic infections remained an independent negative predictor of 90-d mortality (odds ratio 0.14, 95% confidence interval: 0.1-0.4), after controlling for sequential organ failure assessment.

CONCLUSIONS

Urologic infections remained an independent negative predictor of 90-d mortality when compared to other sources of sepsis. Characterization of sepsis source revealed distinct populations and clinical courses, highlighting the importance of understanding different sepsis phenotypes.

IDENTIFICATION OF SUBPHENOTYPES OF SEPSIS-ASSOCIATED LIVER DYSFUNCTION USING CLUSTER ANALYSIS

ABSTRACT

Objectives: We attempted to identify and validate the subphenotypes of sepsis-associated liver dysfunction (SALD) using routine clinical information. Design: This article is a retrospective observational cohort study. Setting: We used the Medical Information Mart for Intensive Care IV database and the eICU Collaborative Research Database. Patients: We included adult patients (age ≥18 years) who developed SALD within the first 48 hours of intensive care unit (ICU) admission. We excluded patients who died or were discharged from the ICU within the first 48 hours of admission. Patients with abnormal liver function before ICU admission were also excluded. Measurements and Main Results: Patients in the MIMIC-IV 1.0 database served as a derivation cohort. Patients in the eICU database were used as validation cohort. We identified four subphenotypes of SALD (subphenotype α, β, γ, δ) using K-means cluster analysis in 5234 patients in derivation cohort. The baseline characteristics and clinical outcomes were compared between the phenotypes using one-way analysis of variance/Kruskal-Wallis test and the χ 2 test. Moreover, we used line charts to illustrate the trend of liver function parameters over 14 days after ICU admission. Subphenotype α (n = 1,055) was the most severe cluster, characterized by shock with multiple organ dysfunction (MODS) group. Subphenotype β (n = 1,179) had the highest median bilirubin level and the highest proportion of patients with underlying liver disease and coexisting coagulopathy (increased bilirubin group). Subphenotype γ (n = 1,661) was the cluster with the highest mean age and had the highest proportion of patients with chronic kidney disease (aged group). Subphenotype δ (n = 1,683) had the lowest 28-day and in-hospital mortality (mild group). The characteristics of clusters in the validation cohort were similar to those in the derivation cohort. In addition, we were surprised to find that GGT levels in subphenotype δ were significantly higher than in other subphenotypes, showing a different pattern from bilirubin. Conclusions: We identified four subphenotypes of SALD that presented with different clinical features and outcomes. These results can provide a valuable reference for understanding the clinical characteristics and associated outcomes to improve the management of patients with SALD in the ICU.

Derivation and Validation of Clinical Phenotypes of the Cardiopulmonary Bypass-Induced Inflammatory Response

BACKGROUND

Precision medicine aims to change treatment from a "one-size-fits-all" approach to customized therapies based on the individual patient. Applying a precision medicine approach to a heterogeneous condition, such as the cardiopulmonary bypass (CPB)-induced inflammatory response, first requires identification of homogeneous subgroups that correlate with biological markers and postoperative outcomes. As a first step, we derived clinical phenotypes of the CPB-induced inflammatory response by identifying patterns in perioperative clinical variables using machine learning and simulation tools. We then evaluated whether these phenotypes were associated with biological response variables and clinical outcomes.

METHODS

This single-center, retrospective cohort study used Cleveland Clinic registry data from patients undergoing cardiac surgery with CPB from January 2010 to March 2020. Biomarker data from a subgroup of patients enrolled in a clinical trial were also included. Patients undergoing emergent surgery, off-pump surgery, transplantation, descending thoracoabdominal aortic surgery, and planned ventricular assist device placement were excluded. Preoperative and intraoperative variables of patient baseline characteristics (demographics, comorbidities, and laboratory data) and perioperative data (procedural data, CPB duration, and hemodynamics) were analyzed to derive clinical phenotypes using K-means-based consensus clustering analysis. Proportion of ambiguously clustered was used to assess cluster size and optimal cluster numbers. After clusters were formed, we summarized perioperative profiles, inflammatory biomarkers (eg, interleukin [IL]-6 and IL-8), kidney biomarkers (eg, urine neutrophil gelatinase-associated lipocalin [NGAL] and IL-18), and clinical outcomes (eg, mortality and hospital length of stay). Pairwise standardized difference was reported for all summarized variables.

RESULTS

Of 36,865 eligible cardiac surgery cases, 25,613 met inclusion criteria. Cluster analysis derived 3 clinical phenotypes: α, β, and γ. Phenotype α (n = 6157 [24%]) included older patients with more comorbidities, including heart and kidney failure. Phenotype β (n = 10,572 [41%]) patients were younger and mostly male. Phenotype γ (n = 8884 [35%]) patients were 58% female and had lower body mass index (BMI). Phenotype α patients had worse outcomes, including longer hospital length of stay (mean = 9 days for α versus 6 for both β [absolute standardized difference {ASD} = 1.15] and γ [ASD = 1.08]), more kidney failure, and higher mortality. Inflammatory biomarkers (IL-6 and IL-8) and kidney injury biomarkers (urine NGAL and IL-18) were higher with the α phenotype compared to β and γ immediately after surgery.

CONCLUSIONS

Deriving clinical phenotypes that correlate with response biomarkers and outcomes represents an initial step toward a precision medicine approach for the management of CPB-induced inflammatory response and lays the groundwork for future investigation, including an evaluation of the heterogeneity of treatment effect.

DJ-1 Deficiency Protects against Sepsis-Induced Myocardial Depression

Oxidative stress is considered one of the early underlying contributors of sepsis-induced myocardial depression. DJ-1, also known as PARK7, has a well-established role as an antioxidant. We have previously shown, in a clinically relevant model of polymicrobial sepsis, DJ-1 deficiency improved survival and bacterial clearance by decreasing ROS production. In the present study, we investigated the role of DJ-1 in sepsis-induced myocardial depression. Here we compared wildtype (WT) with DJ-1 deficient mice at 24 and 48 h after cecal ligation and puncture (CLP). In WT mice, DJ-1 was increased in the myocardium post-CLP. DJ-1 deficient mice, despite enhanced inflammatory and oxidative responses, had an attenuated hypertrophic phenotype, less apoptosis, improved mitochondrial function, and autophagy, that was associated with preservation of myocardial function and improved survival compared to WT mice post-CLP. Collectively, these results identify DJ-1 as a regulator of myocardial function and as such, makes it an attractive therapeutic target in the treatment of early sepsis-induced myocardial depression.

Whole transcriptome analysis to explore the impaired immunological features in critically ill elderly patients with sepsis

BACKGROUND

Sepsis is a frequent complication in critically ill patients, is highly heterogeneous and is associated with high morbidity and mortality rates, especially in the elderly population. Utilizing RNA sequencing (RNA-Seq) to analyze biological pathways is widely used in clinical and molecular genetic studies, but studies in elderly patients with sepsis are still lacking. Hence, we investigated the mortality-relevant biological features and transcriptomic features in elderly patients who were admitted to the intensive care unit (ICU) for sepsis.

METHODS

We enrolled 37 elderly patients with sepsis from the ICU at Taichung Veterans General Hospital. On day-1 and day-8, clinical and laboratory data, as well as blood samples, were collected for RNA-Seq analysis. We identified the dynamic transcriptome and enriched pathways of differentially expressed genes between day-8 and day-1 through DVID enrichment analysis and Gene Set Enrichment Analysis. Then, the diversity of the T cell repertoire was analyzed with MiXCR.

RESULTS

Overall, 37 patients had sepsis, and responders and non-responders were grouped through principal component analysis. Significantly higher SOFA scores at day-7, longer ventilator days, ICU lengths of stay and hospital mortality were found in the non-responder group, than in the responder group. On day-8 in elderly ICU patients with sepsis, genes related to innate immunity and inflammation, such as ZDHCC19, ALOX15, FCER1A, HDC, PRSS33, and PCSK9, were upregulated. The differentially expressed genes (DEGs) were enriched in the regulation of transcription, adaptive immune response, immunoglobulin production, negative regulation of transcription, and immune response. Moreover, there was a higher diversity of T-cell receptors on day-8 in the responder group, than on day-1, indicating that they had better regulated recovery from sepsis compared with the non-response patients.

CONCLUSION

Sepsis mortality and incidence were both high in elderly individuals. We identified mortality-relevant biological features and transcriptomic features with functional pathway and MiXCR analyses based on RNA-Seq data; and found that the responder group had upregulated innate immunity and increased T cell diversity; compared with the non-responder group. RNA-Seq may be able to offer additional complementary information for the accurate and early prediction of treatment outcome.

Mortality and Sequential Organ Failure Assessment Score in Patients With Suspected Sepsis: The Impact of Acute and Preexisting Organ Failures and Infection Likelihood

The Sequential Organ Failure Assessment (SOFA) was chosen in the definition of sepsis due to superior validity in predicting mortality. However, few studies have assessed the contributions of acute versus chronic organ failures to SOFA for mortality prediction.

OBJECTIVES

The main objective in this study was to assess the relative importance of chronic and acute organ failures in mortality prediction in patients with suspected sepsis at hospital admission. We also evaluated how the presence of infection influenced the ability of SOFA to predict 30-day mortality.

DESIGN SETTING AND PARTICIPANTS

Single-center prospective cohort study including 1,313 adult patients with suspected sepsis in rapid response teams in the emergency department.

MAIN OUTCOMES AND MEASURES

The main outcome was 30-day mortality. We measured the maximum total SOFA score during admission (SOFATotal), whereas preexisting chronic organ failure SOFA (SOFAChronic) score was assessed by chart review, allowing calculation of the corresponding acute SOFA (SOFAAcute) score. Likelihood of infection was determined post hoc as "No infection" or "Infection."

RESULTS

SOFAAcute and SOFAChronic were both associated with 30-day mortality, adjusted for age and sex (adjusted odds ratios [AORs], 1.3; 95% CI, 1.3-14 and 1.3; 1.2-1.7), respectively. Presence of infection was associated with lower 30-day mortality (AOR, 0.4; 95% CI, 0.2-0.6), even when corrected for SOFA. In "No infection" patients, SOFAAcute was not associated with mortality (AOR, 1.1; 95% CI, 1.0-1.2), and in this subgroup, neither SOFAAcute greater than or equal to 2 (relative risk [RR], 1.1; 95% CI, 0.6-1.8) nor SOFATotal greater than or equal to 2 (RR, 3.6; 95% CI, 0.9-14.1) was associated with higher mortality.

CONCLUSIONS AND RELEVANCE

Chronic and acute organ failures were equally associated with 30-day mortality in suspected sepsis. A substantial part of the total SOFA score was due to chronic organ failure, calling for caution when using total SOFA in defining sepsis and as an outcome in intervention studies. SOFA's mortality prediction ability was highly dependent on actual presence of infection.

Development of a Reinforcement Learning Algorithm to Optimize Corticosteroid Therapy in Critically Ill Patients with Sepsis

BACKGROUND

The optimal indication, dose, and timing of corticosteroids in sepsis is controversial. Here, we used reinforcement learning to derive the optimal steroid policy in septic patients based on data on 3051 ICU admissions from the AmsterdamUMCdb intensive care database.

METHODS

We identified septic patients according to the 2016 consensus definition. An actor-critic RL algorithm using ICU mortality as a reward signal was developed to determine the optimal treatment policy from time-series data on 277 clinical parameters. We performed off-policy evaluation and testing in independent subsets to assess the algorithm's performance.

RESULTS

Agreement between the RL agent's policy and the actual documented treatment reached 59%. Our RL agent's treatment policy was more restrictive compared to the actual clinician behavior: our algorithm suggested withholding corticosteroids in 62% of the patient states, versus 52% according to the physicians' policy. The 95% lower bound of the expected reward was higher for the RL agent than clinicians' historical decisions. ICU mortality after concordant action in the testing dataset was lower both when corticosteroids had been withheld and when corticosteroids had been prescribed by the virtual agent. The most relevant variables were vital parameters and laboratory values, such as blood pressure, heart rate, leucocyte count, and glycemia.

CONCLUSIONS

Individualized use of corticosteroids in sepsis may result in a mortality benefit, but optimal treatment policy may be more restrictive than the routine clinical practice. Whilst external validation is needed, our study motivates a 'precision-medicine' approach to future prospective controlled trials and practice.

Identification of Distinct Clinical Phenotypes of Heterogeneous Mechanically Ventilated ICU Patients Using Cluster Analysis

This retrospective study aimed to derive the clinical phenotypes of ventilated ICU patients to predict the outcomes on the first day of ventilation. Clinical phenotypes were derived from the eICU Collaborative Research Database (eICU) cohort via cluster analysis and were validated in the Medical Information Mart for Intensive Care (MIMIC-IV) cohort. Four clinical phenotypes were identified and compared in the eICU cohort (n = 15,256). Phenotype A (n = 3112) was associated with respiratory disease, had the lowest 28-day mortality (16%), and had a high extubation success rate (~80%). Phenotype B (n = 3335) was correlated with cardiovascular disease, had the second-highest 28-day mortality (28%), and had the lowest extubation success rate (69%). Phenotype C (n = 3868) was correlated with renal dysfunction, had the highest 28-day mortality (28%), and had the second-lowest extubation success rate (74%). Phenotype D (n = 4941) was associated with neurological and traumatic diseases, had the second-lowest 28-day mortality (22%), and had the highest extubation success rate (>80%). These findings were validated in the validation cohort (n = 10,813). Additionally, these phenotypes responded differently to ventilation strategies in terms of duration of treatment, but had no difference in mortality. The four clinical phenotypes unveiled the heterogeneity of ICU patients and helped to predict the 28-day mortality and the extubation success rate.

Evolution of Clinical Phenotypes of COVID-19 Patients During Intensive Care Treatment: An Unsupervised Machine Learning Analysis

BACKGROUND

Identification of clinical phenotypes in critically ill COVID-19 patients could improve understanding of the disease heterogeneity and enable prognostic and predictive enrichment. However, previous attempts did not take into account temporal dynamics with high granularity. By including the dimension of time, we aim to gain further insights into the heterogeneity of COVID-19.

METHODS

We used granular data from 3202 adult COVID patients in the Dutch Data Warehouse that were admitted to one of 25 Dutch ICUs between February 2020 and March 2021. Parameters including demographics, clinical observations, medications, laboratory values, vital signs, and data from life support devices were selected. Twenty-one datasets were created that each covered 24 h of ICU data for each day of ICU treatment. Clinical phenotypes in each dataset were identified by performing cluster analyses. Both evolution of the clinical phenotypes over time and patient allocation to these clusters over time were tracked.

RESULTS

The final patient cohort consisted of 2438 COVID-19 patients with a ICU mortality outcome. Forty-one parameters were chosen for cluster analysis. On admission, both a mild and a severe clinical phenotype were found. After day 4, the severe phenotype split into an intermediate and a severe phenotype for 11 consecutive days. Heterogeneity between phenotypes appears to be driven by inflammation and dead space ventilation. During the 21-day period, only 8.2% and 4.6% of patients in the initial mild and severe clusters remained assigned to the same phenotype respectively. The clinical phenotype half-life was between 5 and 6 days for the mild and severe phenotypes, and about 3 days for the medium severe phenotype.

CONCLUSIONS

Patients typically do not remain in the same cluster throughout intensive care treatment. This may have important implications for prognostic or predictive enrichment. Prominent dissimilarities between clinical phenotypes are predominantly driven by inflammation and dead space ventilation.

Ultrasound in Sepsis and Septic Shock-From Diagnosis to Treatment

Sepsis and septic shock are among the leading causes of in-hospital mortality worldwide, causing a considerable burden for healthcare. The early identification of sepsis as well as the individuation of the septic focus is pivotal, followed by the prompt initiation of antibiotic therapy, appropriate source control as well as adequate hemodynamic resuscitation. For years now, both emergency department (ED) doctors and intensivists have used ultrasound as an adjunctive tool for the correct diagnosis and treatment of these patients. Our aim was to better understand the state-of-the art role of ultrasound in the diagnosis and treatment of sepsis and septic shock.

METHODS

We conducted an extensive literature search about the topic and reported on the data from the most significant papers over the last 20 years.

RESULTS

We divided each article by topic and exposed the results accordingly, identifying four main aspects: sepsis diagnosis, source control and procedure, fluid resuscitation and hemodynamic optimization, and echocardiography in septic cardiomyopathy.

CONCLUSION

The use of ultrasound throughout the process of the diagnosis and treatment of sepsis and septic shock provides the clinician with an adjunctive tool to better characterize patients and ensure early, aggressive, as well as individualized therapy, when needed. More data are needed to conclude that the use of ultrasound might improve survival in this subset of patients.

Clustering using unsupervised machine learning to stratify the risk of immune-related liver injury

BACKGROUND AND AIM

Immune-related liver injury (liver-irAE) is a clinical problem with a potentially poor prognosis.

METHODS

We retrospectively collected clinical data from patients treated with immune checkpoint inhibitors between September 2014 and December 2021 at the Nagoya University Hospital. Using an unsupervised machine learning method, the Gaussian mixture model, to divide the cohort into clusters based on inflammatory markers, we investigated the cumulative incidence of liver-irAEs in these clusters.

RESULTS

This study included a total of 702 patients. Among them, 492 (70.1%) patients were male, and the mean age was 66.6 years. During the mean follow-up period of 423 days, severe liver-irAEs (Common Terminology Criteria for Adverse Events grade ≥ 3) occurred in 43 patients. Patients were divided into five clusters (a, b, c, d, and e). The cumulative incidence of liver-irAE was higher in cluster c than in cluster a (hazard ratio [HR]: 13.59, 95% confidence interval [CI]: 1.70-108.76, P = 0.014), and overall survival was worse in clusters c and d than in cluster a (HR: 2.83, 95% CI: 1.77-4.50, P < 0.001; HR: 2.87, 95% CI: 1.47-5.60, P = 0.002, respectively). Clusters c and d were characterized by high temperature, C-reactive protein, platelets, and low albumin. However, there were differences in the prevalence of neutrophil count, neutrophil-to-lymphocyte ratio, and liver metastases between both clusters.

CONCLUSIONS

The combined assessment of multiple markers and body temperature may help stratify high-risk groups for developing liver-irAE.

Combination of WFDC2, CHI3L1, and KRT19 in Plasma Defines a Clinically Useful Molecular Phenotype Associated with Prognosis in Critically Ill COVID-19 Patients

BACKGROUND

COVID-19 is now a common disease, but its pathogenesis remains unknown. Blood circulating proteins reflect host defenses against COVID-19. We investigated whether evaluation of longitudinal blood proteomics for COVID-19 and merging with clinical information would allow elucidation of its pathogenesis and develop a useful clinical phenotype.

METHODS

To achieve the first goal (determining key proteins), we derived plasma proteins related to disease severity by using a first discovery cohort. We then assessed the association of the derived proteins with clinical outcome in a second discovery cohort. Finally, the candidates were validated by enzyme-linked immunosorbent assay in a validation cohort to determine key proteins. For the second goal (understanding the associations of the clinical phenotypes with 28-day mortality and clinical outcome), we assessed the associations between clinical phenotypes derived by latent cluster analysis with the key proteins and 28-day mortality and clinical outcome.

RESULTS

We identified four key proteins (WFDC2, GDF15, CHI3L1, and KRT19) involved in critical pathogenesis from the three different cohorts. These key proteins were related to the function of cell adhesion and not immune response. Considering the multicollinearity, three clinical phenotypes based on WFDC2, CHI3L1, and KRT19 were identified that were associated with mortality and clinical outcome.

CONCLUSION

The use of these easily measured key proteins offered new insight into the pathogenesis of COVID-19 and could be useful in a potential clinical application.

LEUKOCYTE PHENOTYPING IN SEPSIS USING OMICS, FUNCTIONAL ANALYSIS, AND IN SILICO MODELING

ABSTRACT

Sepsis is a major health issue and a leading cause of death in hospitals globally. The treatment of sepsis is largely supportive, and there are no therapeutics available that target the underlying pathophysiology of the disease. The development of therapeutics for the treatment of sepsis is hindered by the heterogeneous nature of the disease. The presence of multiple, distinct immune phenotypes ranging from hyperimmune to immunosuppressed can significantly impact the host response to infection. Recently, omics, biomarkers, cell surface protein expression, and immune cell profiles have been used to classify immune status of sepsis patients. However, there has been limited studies of immune cell function during sepsis and even fewer correlating omics and biomarker alterations to functional consequences. In this review, we will discuss how the heterogeneity of sepsis and associated immune cell phenotypes result from changes in the omic makeup of cells and its correlation with leukocyte dysfunction. We will also discuss how emerging techniques such as in silico modeling and machine learning can help in phenotyping sepsis patients leading to precision medicine.

Refining empiric subgroups of pediatric sepsis using machine-learning techniques on observational data

Sepsis contributes to 1 of every 5 deaths globally with 3 million per year occurring in children. To improve clinical outcomes in pediatric sepsis, it is critical to avoid "one-size-fits-all" approaches and to employ a precision medicine approach. To advance a precision medicine approach to pediatric sepsis treatments, this review provides a summary of two phenotyping strategies, empiric and machine-learning-based phenotyping based on multifaceted data underlying the complex pediatric sepsis pathobiology. Although empiric and machine-learning-based phenotypes help clinicians accelerate the diagnosis and treatments, neither empiric nor machine-learning-based phenotypes fully encapsulate all aspects of pediatric sepsis heterogeneity. To facilitate accurate delineations of pediatric sepsis phenotypes for precision medicine approach, methodological steps and challenges are further highlighted.

Changes in Heparin-Binding Protein, Procalcitonin, and C-Reactive Protein Within the First 72 Hours Predict 28-Day Mortality in Patients Admitted to the Intensive Care Unit with Septic Shock

BACKGROUND This study aimed to evaluate the possible associations of heparin-binding protein (HBP), procalcitonin (PCT), and C-reactive protein (CRP) levels with 28-day mortality in septic shock patients admitted to Intensive Care Units (ICUs). MATERIAL AND METHODS Blood samples were taken at ICU admission and measured again 72 h later to calculate changes in HBP (ΔHBP), changes in PCT (ΔPCT), changes in CRP (ΔCRP), and changes in Sequential Organ Failure Assessment (DSOFA) relative to baseline. RESULTS Variables included in the univariable logistic regression model for survival were age, Acute Physiology and Chronic Health Evaluation (APACHE) II scores, decreasing ΔSOFA, decreasing DHBP, decreasing ΔPCT, and decreasing ΔCRP. Survival was directly related to decreasing ΔHBP with odds ratio (OR)=9.95 (95% confidence interval [CI] 4.63 to 21.35; P<0.001), decreasing ΔPCT with OR=7.85 (3.74 to 16.49; P<0.001), decreasing ΔCRP with OR=5.83 (2.84 to 11.97; P<0.001), decreasing ΔSOFA with OR=1.93 (1.00 to 3.75; P=0.051) and APACHE II score with OR=1.93 (1.14 to 1.68; P=0.001). In a multivariable logistic regression model for survival, only decreasing DHBP with OR=7.18 (2.91 to 17.69; P<0.001), decreasing ΔPCT with OR=5.17 (2.12 to 12.56; P<0.001), and decreasing ΔCRP with OR=4.33 (1.77 to 10.61; P=0.001) remained significant. CONCLUSIONS Measuring changes in HBP, PCT, and CRP within 72 h of admission may aid in predicting 28-day mortality for patients with septic shock in ICUs.

Research frontiers and trends in the application of artificial intelligence to sepsis: A bibliometric analysis

Background

With the increasing interest of academics in the application of artificial intelligence to sepsis, thousands of papers on this field had been published in the past few decades. It is difficult for researchers to understand the themes and latest research frontiers in this field from a multi-dimensional perspective. Consequently, the purpose of this study is to analyze the relevant literature in the application of artificial intelligence to sepsis through bibliometrics software, so as to better understand the development status, study the core hotspots and future development trends of this field.

Methods

We collected relevant publications in the application of artificial intelligence to sepsis from the Web of Science Core Collection in 2000 to 2021. The type of publication was limited to articles and reviews, and language was limited to English. Research cooperation network, journals, cited references, keywords in this field were visually analyzed by using CiteSpace, VOSviewer, and COOC software.

Results

A total of 8,481 publications in the application of artificial intelligence to sepsis between 2000 and 2021 were included, involving 8,132 articles and 349 reviews. Over the past 22 years, the annual number of publications had gradually increased exponentially. The USA was the most productive country, followed by China. Harvard University, Schuetz, Philipp, and Intensive Care Medicine were the most productive institution, author, and journal, respectively. Vincent, Jl and Critical Care Medicine were the most cited author and cited journal, respectively. Several conclusions can be drawn from the analysis of the cited references, including the following: screening and identification of sepsis biomarkers, treatment and related complications of sepsis, and precise treatment of sepsis. Moreover, there were a spike in searches relating to machine learning, antibiotic resistance and accuracy based on burst detection analysis.

Conclusion

This study conducted a comprehensive and objective analysis of the publications on the application of artificial intelligence in sepsis. It can be predicted that precise treatment of sepsis through machine learning technology is still research hotspot in this field.

Modeling sepsis, with a special focus on large animal models of porcine peritonitis and bacteremia

Infectious diseases, which often result in deadly sepsis or septic shock, represent a major global health problem. For understanding the pathophysiology of sepsis and developing new treatment strategies, reliable and clinically relevant animal models of the disease are necessary. In this review, two large animal (porcine) models of sepsis induced by either peritonitis or bacteremia are introduced and their strong and weak points are discussed in the context of clinical relevance and other animal models of sepsis, with a special focus on cardiovascular and immune systems, experimental design, and monitoring. Especially for testing new therapeutic strategies, the large animal (porcine) models represent a more clinically relevant alternative to small animal models, and the findings obtained in small animal (transgenic) models should be verified in these clinically relevant large animal models before translation to the clinical level.

Temporal Supervised Contrastive Learning for Modeling Patient Risk Progression

We consider the problem of predicting how the likelihood of an outcome of interest for a patient changes over time as we observe more of the patient's data. To solve this problem, we propose a supervised contrastive learning framework that learns an embedding representation for each time step of a patient time series. Our framework learns the embedding space to have the following properties: (1) nearby points in the embedding space have similar predicted class probabilities, (2) adjacent time steps of the same time series map to nearby points in the embedding space, and (3) time steps with very different raw feature vectors map to far apart regions of the embedding space. To achieve property (3), we employ a nearest neighbor pairing mechanism in the raw feature space. This mechanism also serves as an alternative to "data augmentation", a key ingredient of contrastive learning, which lacks a standard procedure that is adequately realistic for clinical tabular data, to our knowledge. We demonstrate that our approach outperforms state-of-the-art baselines in predicting mortality of septic patients (MIMIC-III dataset) and tracking progression of cognitive impairment (ADNI dataset). Our method also consistently recovers the correct synthetic dataset embedding structure across experiments, a feat not achieved by baselines. Our ablation experiments show the pivotal role of our nearest neighbor pairing.

[Pathophysiology of sepsis]

Up to now, sepsis is one of the most threatening diseases and its therapy remains challenging. Sepsis is currently defined as a severely dysregulated immune response to an infection resulting in organ dysfunction. The pathophysiology is mainly driven by exogenous PAMPs ("pathogen-associated molecular patterns") and endogenous DAMPs ("damage-associated molecular patterns"), which can activate PRRs ("pattern recognition receptors") on different cell types (mainly immune cells), leading to the initiation of manifold downstream pathways and a perpetuation of patients' immune response. Sepsis is neither an exclusive pro- nor an anti-inflammatory disease: both processes take place in parallel, resulting in an individual immunologic disease state depending on the severity of each component at different time points. Septic shock is a complex disorder of the macro- and microcirculation, provoking a severe lack of oxygenation further aggravating sepsis defining organ dysfunctions. An in-depth knowledge of the heterogeneity and the time-dependency of the septic immunopathology will be essential for the design of future sepsis trials and therapy planning in patients with sepsis. The big aim is to achieve a more individualized treatment strategy in patients suffering from sepsis or septic shock.

Prognostic Value of Antithrombin Activity Levels in the Early Phase of Intensive Care: A 2-Center Retrospective Cohort Study

To investigate the relationship between antithrombin (AT) activity level and prognosis in patients requiring intensive care. Patients whose AT activity was measured within 24 h of intensive care unit (ICU) admission were enrolled for analysis. The primary endpoint was mortality at discharge. Prognostic accuracy was examined using receiver operating characteristic (ROC) curves and cox hazard regression analysis. Patients were divided into 6 groups based on predicted mortality, and a χ2 independence test was performed on the prognostic value of AT activity for each predicted mortality; P < .05 was considered significant. A total of 281 cases were analyzed. AT activity was associated with mortality at discharge (AT% [interquartile range, IQR]): survivor group, 69 (56-86) versus nonsurvivor group, 56 (44-73), P = .0003). We found an increasing risk for mortality in both the lowest level of AT activity (<50%; hazard ratio [HR] 2.43, 95% confidence interval [CI] 1.20-4.89, P = .01) and the middle-low level of AT activity (≥ 50% and < 70%; HR 2.06, 95% CI 1.06-4.02, P = .03), compared with the normal AT activity level (≥ 70%). ROC curve analysis showed that the prediction accuracy of AT was an area under the curve (AUC) of 0.66 (cutoff 58%, sensitivity 61.4%, specificity 68.2%, P = .0003). AT activity was significantly prognostic in the group with 20% to 50% predicted mortality (AUC 0.74, sensitivity: 24.0%-55.5%, specificity: 83.3%-93.0%). An early decrease in AT activity level in ICU patients may be a predictor of mortality at discharge.

Sepsis-induced disseminated intravascular coagulation: an international estrangement of disease concept

Disseminated intravascular coagulation (DIC) is an acquired syndrome characterized by widespread intravascular activation of coagulation, which can be caused by infectious and noninfectious insults, such as trauma, postcardiac arrest syndrome, and malignant diseases. At present, diagnosis and treatment of DIC clearly differ between Japan and Western countries; in Japan, DIC has long been considered a therapeutic target, and much evidence on DIC has been published. However, there has recently been no international consensus on whether DIC should be a therapeutic target with anticoagulant therapy. This review describes the coagulofibrinolytic system abnormalities associated with sepsis and discusses related management strategies. It also explores the reasons why DIC is perceived differently in different regions. There is a major discrepancy between diagnostic and treatment options in Japan, which are based on holistic assessments of trials, as well as the results of post hoc subgroup analyses and observational studies, and those in Western countries, which are based mainly on the results of sepsis mega trials, especially randomized controlled trials. The differences might also be due to various patient factors in each region, especially racial characteristics in thrombolytic mechanisms, and differences in interpretation of evidence for candidate drugs. Hence, Japanese researchers need to distribute their high-quality clinical research data not only to Japan but also to the rest of the world.

Endotoxemic Shock: A Molecular Phenotype in Sepsis

BACKGROUND

Marked heterogeneity exists among patients with sepsis, both in terms of distribution of organ dysfunction and its severity. Such heterogeneity could be explained by the presence of multiple subtypes of sepsis that may have important implications for treatment.

METHODS

Narrative review of published literature involving endotoxin from 1970 to 2022.

RESULTS

In humans, endotoxemia is most consistently associated with a specific pattern of organ failure including shock, endothelial dysfunction, acute kidney injury, and hepatic dysfunction. This pattern is consistent with complement activation and uncontrolled inflammation, two features of endotoxemia. Unbiased discovery using artificial intelligence also identifies a subtype of sepsis which features these same organ failures.

CONCLUSION

Endotoxin appears to represent an important molecular phenotype of sepsis with unique clinical features and high mortality.

Machine-learning-derived sepsis bundle of care

PURPOSE

Compliance to the Surviving Sepsis Campaign (SSC) guidelines is limited. This is known to be associated with increased mortality. The aim of this retrospective cohort study was to identify among the SCC guidelines the optimal bundle of recommendations that minimize 28-day mortality.

METHODS

We used a training cohort to identify, using a least absolute shrinkage and selection operator penalized machine learning model, this bundle. Patients with sepsis/septic shock admitted to the intensive care unit (ICU) were extracted from two US databases, the Medical Information Mart for Intensive Care-IV (MIMIC-IV) database (training and internal validation cohorts) and the eICU Collaborative Research Database (eICU-CRD) (external validation cohort). In the validation cohorts, we defined a bundle group that includes patients who were treated with at least all the recommendations selected in our bundle and a no-bundle group that includes patients in whom at least one recommendation from our bundle was omitted.

RESULTS

All-cause 28-day mortality was the primary outcome measure. A total of 42,735 patients were included. Six recommendations (antimicrobials, balanced crystalloid, insulin therapy, corticosteroids, vasopressin, and bicarbonate therapy) were identified from the training cohort to be included in our bundle. In the propensity score-(PS)-matched internal validation cohort, the bundle group was associated with a lower mortality (OR 0.41 [0.33-0.53]; p < 0.001) compared to the no-bundle group. This was confirmed in the PS-matched external validation cohort (OR 0.75 [0.60-0.94]; p 0.02).

CONCLUSION

Our bundle of six recommendations is associated with a dramatic reduction in mortality in sepsis and septic shock. This bundle needs to be evaluated prospectively.

A framework for understanding, designing, developing and evaluating learning health systems

Introduction

A Learning Health System is not a technical project. It is the evolution of an existing health system into one capable of learning from every patient. This paper outlines a recently published framework intended to aid the understanding, design, development and evaluation of Learning Health Systems.

Methods

This work extended an existing repository of Learning Health System evidence, adding five more workshops. The total was subjected to thematic analysis, yielding a framework of elements important to understanding, designing, developing and evaluating Learning Health Systems. Purposeful literature reviews were conducted on each element. The findings were revised following a review by a group of international experts.

Results

The resulting framework was arranged around four questions:What is our rationale for developing a Learning Health System?There can be many reasons for developing a Learning Health System. Understanding these will guide its development.What sources of complexity exist at the system and the intervention level?An understanding of complexity is central to making Learning Health Systems work. The non-adoption, abandonment, scale-up, spread and sustainability framework was utilised to help understand and manage it.What strategic approaches to change do we need to consider?A range of strategic issues must be addressed to enable successful change in a Learning Health System. These include, strategy, organisational structure, culture, workforce, implementation science, behaviour change, co-design and evaluation.What technical building blocks will we need?A Learning Health System must capture data from practice, turn it into knowledge and apply it back into practice. There are many methods to achieve this and a range of platforms to help.

Discussion

The results form a framework for understanding, designing, developing and evaluating Learning Health Systems at any scale.

Conclusion

It is hoped that this framework will evolve with use and feedback.

Association between Pre-Existing Long-Term β-Blocker Therapy and the Outcomes of Sepsis-Associated Coagulopathy: A Retrospective Study

Background and Objectives: Previous studies have suggested that long-term β-blocker therapy before sepsis is associated with reduced mortality. Sepsis-associated coagulopathy (SAC) remains a common complication in patients with sepsis and is associated with increased mortality. Adrenergic pathways are involved in the regulation of the coagulation system. Pre-existing long-term β-blocker therapy may have potentially beneficial effects on SAC and has yet to be well characterized. We aimed to assess the potential association between pre-existing long-term β-blocker therapy and the outcomes of patients with SAC. Materials and Methods: This study retrospectively screened the clinical data of adult patients with SAC admitted to the Intensive Care Unit (ICU) and respiratory ICU between May 2020 and October 2022. Patients with SAC who took any β-blocker for at least one year were considered pre-existing long-term β-blocker therapy. All enrolled patients were followed up for 28 days or until death. Results: Among the 228 SAC patients, 48 received long-term β-blocker therapy before septic episodes. Pre-existing long-term β-blocker therapy was associated with reduced vasopressor requirements and a decreased 28-day mortality (log-rank test: p = 0.041). In particular, long-term β-blocker therapy was related to substantially lower D-dimer levels and a trend of improved activated partial thromboplastin time in patients with SAC during initial ICU admission. Multivariable regression analysis showed that long-term β-blocker therapy was significantly and independently associated with a 28-day mortality among patients with SAC (adjusted odds ratio, 0.55; 95% confidence interval, (0.32-0.94); p = 0.030). Conclusions: Pre-existing long-term β-blocker therapy might be associated with reduced vasopressor requirements and a decreased 28-day mortality among patients with SAC, providing evidence for the protective effect of β-blockers against SAC in managing sepsis.

Using machine learning for the early prediction of sepsis-associated ARDS in the ICU and identification of clinical phenotypes with differential responses to treatment

Background: An early diagnosis model with clinical phenotype classification is key for the early identification and precise treatment of sepsis-associated acute respiratory distress syndrome (ARDS). This study aimed to: 1) build a machine learning diagnostic model for patients with sepsis-associated ARDS using easily accessible early clinical indicators, 2) conduct rapid classification of clinical phenotypes in this population, and 3) explore the differences in clinical characteristics, outcomes, and treatment responses of different phenotypes. Methods: This study is based on data from the Telehealth Intensive Care Unit (eICU) and Medical Information Mart for Intensive Care IV (MIMIC-IV). We trained and tested the early diagnostic model of sepsis-associated ARDS patients in the eICU. We used key predictive indicators to cluster sepsis-associated ARDS patients and determine the characteristics and clinical outcomes of different phenotypes, as well to explore the differences of in-hospital mortality among different the positive end-expiratory pressure (PEEP) levels in different phenotypes. These results are verified in MIMIC-IV to evaluate whether they are repeatable. Results: Among the diagnostic models constructed in 19,249 sepsis patients and 5,947 sepsis-associated ARDS patients, the AdaBoost (Decision Tree) model achieved the best performance with an area under the receiver operating characteristic curve (AUC) of 0.895, which is higher than that of the traditional Logistic Regression model (Z = -2.40,p = 0.013), and the accuracy of 70.06%, sensitivity of 78.11% and specificity of 78.74%. We simultaneously identified three sepsis-associated ARDS phenotypes. Cluster 0 (n = 3,669) had the lowest in-hospital mortality rate (6.51%) and fewer laboratory abnormalities (lower WBC (median:15.000 K/mcL), lower blood glucose (median:158.000 mg/dl), lower creatinine (median:1.200 mg/dl), lower lactic acid (median:3.000 mmol/L); p < 0.001). Cluster 1 (n = 1,554) had the highest in-hospital mortality rate (75.29%) and the most laboratory abnormalities (higher WBC (median:18.300 K/mcL), higher blood glucose (median:188.000 mg/dl), higher creatinine (median:2.300 mg/dl), higher lactic acid (median:3.900 mmol/L); p < 0.001). Cluster 2 (n = 724) had the most complex condition, with a moderate in-hospital mortality rate (29.7%) and the longest intensive care unit stay. In Clusters 0 and 1, patients with high PEEP had higher in-hospital mortality rate than those with low PEEP, but the opposite trend was seen in Cluster 2. These results were repeatable in 11,935 patients with sepsis and 2,699 patients with sepsis-associated ARDS patients in the MIMIC-IV cohort. Conclusion: A machine learning diagnostic model of sepsis-associated ARDS patients was established. Three phenotypes with different clinical features and outcomes were clustered, and these had different therapeutic responses. These findings are helpful for the early and rapid identification of sepsis-associated ARDS patients and their precise individualized treatment.

Proteomic changes associated with racial background and sepsis survival outcomes

Intra-abdominal infection is a common cause of sepsis, and intra-abdominal sepsis leads to ∼156 000 U.S. deaths annually. African American/Black adults have higher incidence and mortality rates from sepsis compared to Non-Hispanic White adults. A limited number of studies have traced survival outcomes to molecular changes; however, these studies primarily only included Non-Hispanic White adults. Our goal is to better understand molecular changes that may contribute to differences in sepsis survival in African American/Black and Non-Hispanic White adults with primary intra-abdominal infection. We employed discovery-based plasma proteomics of patient samples from the Protocolized Care for Early Septic Shock (ProCESS) cohort (N = 107). We identified 49 proteins involved in the acute phase response and complement system whose expression levels are associated with both survival outcome and racial background. Additionally, 82 proteins differentially-expressed in survivors were specific to African American/Black or Non-Hispanic White patients, suggesting molecular-level heterogeneity in sepsis patients in key inflammatory pathways. A smaller, robust set of 19 proteins were in common in African American/Black and Non-Hispanic White survivors and may represent potential universal molecular changes in sepsis. Overall, this study identifies molecular factors that may contribute to differences in survival outcomes in African American/Black patients that are not fully explained by socioeconomic or other non-biological factors.

Sepsis biomarkers and diagnostic tools with a focus on machine learning

Over the last years, there have been advances in the use of data-driven techniques to improve the definition, early recognition, subtypes characterisation, prognostication and treatment personalisation of sepsis. Some of those involve the discovery or evaluation of biomarkers or digital signatures of sepsis or sepsis sub-phenotypes. It is hoped that their identification may improve timeliness and accuracy of diagnosis, suggest physiological pathways and therapeutic targets, inform targeted recruitment into clinical trials, and optimise clinical management. Given the complexities of the sepsis response, panels of biomarkers or models combining biomarkers and clinical data are necessary, as well as specific data analysis methods, which broadly fall under the scope of machine learning. This narrative review gives a brief overview of the main machine learning techniques (mainly in the realms of supervised and unsupervised methods) and published applications that have been used to create sepsis diagnostic tools and identify biomarkers.

Gaps and opportunities in sepsis translational research

Infection initiates sepsis, but the clinical disease arises through the innate immune response of the host. A rapidly evolving understanding of the biology of that response has not been paralleled by the development of successful new treatment. The COVID-19 pandemic has begun to change this revealing the promise of distinct therapeutic approaches and the feasibility of new approaches to evaluate them. We review the history of mediator-targeted therapy for sepsis and explore the conceptual, biological, technological, and organizational challenges that must be addressed to enable the development of effective treatments for a leading cause of global morbidity and mortality.

Host Gene Expression to Predict Sepsis Progression

OBJECTIVES

Sepsis causes significant mortality. However, most patients who die of sepsis do not present with severe infection, hampering efforts to deliver early, aggressive therapy. It is also known that the host gene expression response to infection precedes clinical illness. This study seeks to develop transcriptomic models to predict progression to sepsis or shock within 72 hours of hospitalization and to validate previously identified transcriptomic signatures in the prediction of 28-day mortality.

DESIGN

Retrospective differential gene expression analysis and predictive modeling using RNA sequencing data.

PATIENTS

Two hundred seventy-seven patients enrolled at four large academic medical centers; all with clinically adjudicated infection were considered for inclusion in this study.

MEASUREMENTS AND MAIN RESULTS

Sepsis progression was defined as an increase in Sepsis 3 category within 72 hours. Transcriptomic data were generated using RNAseq of whole blood. Least absolute shrinkage and selection operator modeling was used to identify predictive signatures for various measures of disease progression. Four previously identified gene signatures were tested for their ability to predict 28-day mortality. There were no significant differentially expressed genes in 136 subjects with worsened Sepsis 3 category compared with 141 nonprogressor controls. There were 1,178 differentially expressed genes identified when sepsis progression was defined as ICU admission or 28-day mortality. A model based on these genes predicted progression with an area under the curve of 0.71. Validation of previously identified gene signatures to predict sepsis mortality revealed area under the receiver operating characteristic values of 0.70-0.75 and no significant difference between signatures.

CONCLUSIONS

Host gene expression was unable to predict sepsis progression when defined by an increase in Sepsis-3 category, suggesting this definition is not a useful framework for transcriptomic prediction methods. However, there was a differential response when progression was defined as ICU admission or death. Validation of previously described signatures predicted 28-day mortality with insufficient accuracy to offer meaningful clinical utility.

Timing and Spectrum of Antibiotic Treatment for Suspected Sepsis and Septic Shock: Why so Controversial?

Sepsis guidelines and mandates encourage increasingly aggressive time-to-antibiotic targets for broad-spectrum antimicrobials for suspected sepsis and septic shock. This has caused considerable controversy due to weaknesses in the underlying evidence and fear that overly strict antibiotic deadlines may harm patients by perpetuating or escalating overtreatment. Indeed, a third or more of patients currently treated for sepsis and septic shock have noninfectious or nonbacterial conditions. These patients risk all the potential harms of antibiotics without their possible benefits. Updated Surviving Sepsis Campaign guidelines now emphasize the importance of tailoring antibiotics to each patient's likelihood of infection, risk for drug-resistant pathogens, and severity-of-illness.

Clustering analysis of geriatric and acute characteristics in a cohort of very old patients on admission to ICU

PURPOSE

The biological and functional heterogeneity in very old patients constitutes a major challenge to prognostication and patient management in intensive care units (ICUs). In addition to the characteristics of acute diseases, geriatric conditions such as frailty, multimorbidity, cognitive impairment and functional disabilities were shown to influence outcome in that population. The goal of this study was to identify new and robust phenotypes based on the combination of these features to facilitate early outcome prediction.

METHODS

Patients aged 80 years old or older with and without limitations of life-sustaining treatment and with complete data were recruited from the VIP2 study for phenotyping and from the COVIP study for external validation. The sequential organ failure assessment (SOFA) score and its sub-scores taken on admission to ICU as well as demographic and geriatric patient characteristics were subjected to clustering analysis. Phenotypes were identified after repeated bootstrapping and clustering runs.

RESULTS

In patients from the VIP2 study without limitations of life-sustaining treatment (n = 1977), ICU mortality was 12% and 30-day mortality 19%. Seven phenotypes with distinct profiles of acute and geriatric characteristics were identified in that cohort. Phenotype-specific mortality within 30 days ranged from 3 to 57%. Among the patients assigned to a phenotype with pronounced geriatric features and high SOFA scores, 50% died in ICU and 57% within 30 days. Mortality differences between phenotypes were confirmed in the COVIP study cohort (n = 280).

CONCLUSIONS

Phenotyping of very old patients on admission to ICU revealed new phenotypes with different mortality and potential need for anticipatory intervention.

Derivation and Validation of Vasoactive Inotrope Score Trajectory Groups in Critically Ill Children With Shock

OBJECTIVES

To determine whether there are clinically relevant and reproducible Vasoactive Inotrope Score (VIS) trajectories in children with shock during the acute phase of critical illness.

DESIGN

Retrospective, observational cohort study.

SETTING

Two tertiary, academic PICUs.

PATIENTS

Children (< 18 yr old) who required vasoactive infusions within 24 hours of admission to the PICU. Those admitted post cardiac surgery were excluded.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

An hourly VIS was calculated for the first 72 hours after initiation of vasoactives. Group-based trajectory modeling (GBTM) was applied to a derivation set (75% of encounters) and compared with the trajectories in a validation set (25% of encounters) using the same variables. The primary outcome was in-hospital mortality, and the secondary outcome was multiple organ dysfunction syndrome (MODS) on day 7. A total of 1,828 patients met inclusion criteria, and 309 (16.9%) died. GBTM identified four subgroups that were reproducible in the validation set: "Mild, fast resolving shock" ( n = 853 [47%]; mortality 9%), "Moderate, slow resolving shock" ( n = 422 [23%]; mortality 15%), "Moderate, prolonged shock" ( n = 312 [17%]; mortality 21%), and "Severe, prolonged shock" ( n = 241 [13%]; mortality 40%). There was a significant difference in mortality, MODS on day 7, and suspected infection ( p < 0.001) across groups. The "Mild, fast resolving shock" and "Severe, prolonged shock" groups were identifiable within the first 24 hours. The "Moderate, slow resolving" and "Moderate, prolonged shock" groups were indistinguishable in the first 24 hours after initiation of vasoactives but differed in in-hospital mortality and MODS on day 7. Hydrocortisone administration was independently associated with poor outcomes in the "Mild, fast resolving shock" group.

CONCLUSIONS

We uncovered four distinct and reproducible VIS trajectory groups that were associated with different risk factors, response to therapy, and outcomes in children with shock. Characterizing VIS trajectory groups in the acute phase of critical illness may enable better prognostication and more targeted management.

Validating a Proteomic Signature of Severe COVID-19

COVID-19 is a heterogenous disease. Biomarker-based approaches may identify patients at risk for severe disease, who may be more likely to benefit from specific therapies. Our objective was to identify and validate a plasma protein signature for severe COVID-19.

DESIGN

Prospective observational cohort study.

SETTING

Two hospitals in the United States.

PATIENTS

One hundred sixty-seven hospitalized adults with COVID-19.

INTERVENTION

None.

MEASUREMENTS AND MAIN RESULTS

We measured 713 plasma proteins in 167 hospitalized patients with COVID-19 using a high-throughput platform. We classified patients as nonsevere versus severe COVID-19, defined as the need for high-flow nasal cannula, mechanical ventilation, extracorporeal membrane oxygenation, or death, at study entry and in 7-day intervals thereafter. We compared proteins measured at baseline between these two groups by logistic regression adjusting for age, sex, symptom duration, and comorbidities. We used lead proteins from dysregulated pathways as inputs for elastic net logistic regression to identify a parsimonious signature of severe disease and validated this signature in an external COVID-19 dataset. We tested whether the association between corticosteroid use and mortality varied by protein signature. One hundred ninety-four proteins were associated with severe COVID-19 at the time of hospital admission. Pathway analysis identified multiple pathways associated with inflammatory response and tissue repair programs. Elastic net logistic regression yielded a 14-protein signature that discriminated 90-day mortality in an external cohort with an area under the receiver-operator characteristic curve of 0.92 (95% CI, 0.88-0.95). Classifying patients based on the predicted risk from the signature identified a heterogeneous response to treatment with corticosteroids (p = 0.006).

CONCLUSIONS

Inpatients with COVID-19 express heterogeneous patterns of plasma proteins. We propose a 14-protein signature of disease severity that may have value in developing precision medicine approaches for COVID-19 pneumonia.

Identifying inflammatory phenotypes to target mechanism-specific treatments in sepsis

A clinical trial by Leventogiannis et al.¹ suggests that ferritin and HLA-DR monocyte receptor expression can identify septic patients with macrophage-activation-like syndrome (MALS), or immunoparalysis, and that targeting IL-1ra treatment with this strategy may improve outcomes.

Artificial intelligence for precision medicine in autoimmune liver disease

Autoimmune liver diseases (AiLDs) are rare autoimmune conditions of the liver and the biliary tree with unknown etiology and limited treatment options. AiLDs are inherently characterized by a high degree of complexity, which poses great challenges in understanding their etiopathogenesis, developing novel biomarkers and risk-stratification tools, and, eventually, generating new drugs. Artificial intelligence (AI) is considered one of the best candidates to support researchers and clinicians in making sense of biological complexity. In this review, we offer a primer on AI and machine learning for clinicians, and discuss recent available literature on its applications in medicine and more specifically how it can help to tackle major unmet needs in AiLDs.

Remdesivir for Patients Hospitalized with COVID-19 Severe Pneumonia: A National Cohort Study (Remdeco-19)

BACKGROUND

Given the rapidly evolving pandemic of COVID-19 in 2020, authorities focused on the repurposing of available drugs to develop timely and cost-effective therapeutic strategies. Evidence suggested the potential utility of remdesivir in the framework of an early access program. REMDECO-19 is a multicenter national cohort study assessing the ability of remdesivir to improve the outcome of patients hospitalized with COVID-19.

METHODS

We conducted a retrospective real-life study that included all patients from the early access program of remdesivir in France. The primary endpoint was the clinical course evolution of critically ill and hospitalized COVID-19 patients treated with remdesivir. Secondary endpoints were the SOFA score evolution within 29 days following the admission and mortality at 29 and 90 days.

RESULTS

Eighty-five patients were enrolled in 22 sites from January to April 2020. The median WHO and SOFA scores were respectively reduced by two and six points between days 1 and 29. Improvement in the WHO-CPS and the SOFA score were observed in 83.5% and 79.3% of patients, respectively, from day 10. However, there was no effect of remdesivir on the 90-day survival based on the control cohort for hospitalized COVID-19 patients with invasive ventilation.

CONCLUSIONS

SOFA score appeared to be an attractive approach to assess remdesivir efficacy and stratify its utilization or not in critically ill patients with COVID-19. This study brings a new clinical benchmark for therapeutic decision making and supports the use of remdesivir for some hospitalized COVID-19 patients.

An immune dysfunction score for stratification of patients with acute infection based on whole-blood gene expression

Dysregulated host responses to infection can lead to organ dysfunction and sepsis, causing millions of global deaths each year. To alleviate this burden, improved prognostication and biomarkers of response are urgently needed. We investigated the use of whole-blood transcriptomics for stratification of patients with severe infection by integrating data from 3149 samples from patients with sepsis due to community-acquired pneumonia or fecal peritonitis admitted to intensive care and healthy individuals into a gene expression reference map. We used this map to derive a quantitative sepsis response signature (SRSq) score reflective of immune dysfunction and predictive of clinical outcomes, which can be estimated using a 7- or 12-gene signature. Last, we built a machine learning framework, SepstratifieR, to deploy SRSq in adult and pediatric bacterial and viral sepsis, H1N1 influenza, and COVID-19, demonstrating clinically relevant stratification across diseases and revealing some of the physiological alterations linking immune dysregulation to mortality. Our method enables early identification of individuals with dysfunctional immune profiles, bringing us closer to precision medicine in infection.

Septic Shock: Phenotypes and Outcomes

INTRODUCTION

Sepsis is a heterogeneous syndrome that results in life-threatening organ dysfunction. Our goal was to determine the relevant variables and patient phenotypes to use in predicting sepsis outcomes.

METHODS

We performed an ancillary study concerning 119 patients with septic shock at intensive care unit (ICU) admittance (T0). We defined clinical worsening as having an increased sequential organ failure assessment (SOFA) score of ≥ 1, 48 h after admission (ΔSOFA ≥ 1). We performed univariate and multivariate analyses based on the 28-day mortality rate and ΔSOFA ≥ 1 and determined three patient phenotypes: safe, intermediate and unsafe. The persistence of the intermediate and unsafe phenotypes after T0 was defined as a poor outcome.

RESULTS

At T0, the multivariate analysis showed two variables associated with 28-day mortality rate: norepinephrine dose and serum lactate concentration. Regarding ΔSOFA ≥ 1, we identified three variables at T0: norepinephrine dose, lactate concentration and venous-to-arterial carbon dioxide difference (P(v-a)CO₂). At T0, the three phenotypes (safe, intermediate and unsafe) were found in 28 (24%), 70 (59%) and 21 (18%) patients, respectively. We thus suggested using an algorithm featuring norepinephrine dose, lactate concentration and P(v-a)CO₂ to predict patient outcomes and obtained an area under the curve (AUC) of 74% (63-85%).

CONCLUSION

Our findings highlight the fact that identifying relevant variables and phenotypes may help physicians predict patient outcomes.

Development and validation of novel sepsis subphenotypes using trajectories of vital signs

PURPOSE

Sepsis is a heterogeneous syndrome and identification of sub-phenotypes is essential. This study used trajectories of vital signs to develop and validate sub-phenotypes and investigated the interaction of sub-phenotypes with treatment using randomized controlled trial data.

METHODS

All patients with suspected infection admitted to four academic hospitals in Emory Healthcare between 2014-2017 (training cohort) and 2018-2019 (validation cohort) were included. Group-based trajectory modeling was applied to vital signs from the first 8 h of hospitalization to develop and validate vitals trajectory sub-phenotypes. The associations between sub-phenotypes and outcomes were evaluated in patients with sepsis. The interaction between sub-phenotype and treatment with balanced crystalloids versus saline was tested in a secondary analysis of SMART (Isotonic Solutions and Major Adverse Renal Events Trial).

RESULTS

There were 12,473 patients with suspected infection in training and 8256 patients in validation cohorts, and 4 vitals trajectory sub-phenotypes were found. Group A (N = 3483, 28%) were hyperthermic, tachycardic, tachypneic, and hypotensive. Group B (N = 1578, 13%) were hyperthermic, tachycardic, tachypneic (not as pronounced as Group A) and hypertensive. Groups C (N = 4044, 32%) and D (N = 3368, 27%) had lower temperatures, heart rates, and respiratory rates, with Group C normotensive and Group D hypotensive. In the 6,919 patients with sepsis, Groups A and B were younger while Groups C and D were older. Group A had the lowest prevalence of congestive heart failure, hypertension, diabetes mellitus, and chronic kidney disease, while Group B had the highest prevalence. Groups A and D had the highest vasopressor use (p < 0.001 for all analyses above). In logistic regression, 30-day mortality was significantly higher in Groups A and D (p < 0.001 and p = 0.03, respectively). In the SMART trial, sub-phenotype significantly modified treatment effect (p = 0.03). Group D had significantly lower odds of mortality with balanced crystalloids compared to saline (odds ratio (OR) 0.39, 95% confidence interval (CI) 0.23-0.67, p < 0.001).

CONCLUSION

Sepsis sub-phenotypes based on vital sign trajectory were consistent across cohorts, had distinct outcomes, and different responses to treatment with balanced crystalloids versus saline.

Evidence-based updates to the 2021 Surviving Sepsis Campaign guidelines: Part 1: Background, pathophysiology, and emerging treatments

ABSTRACT

Sepsis identification and treatment has changed significantly over the last few decades. Despite this, sepsis is still associated with significant morbidity and mortality. This first of a two-part series reviews the history of modern sepsis and presents new research in pathophysiology, treatment, and postsepsis care.

Selenocompounds and Sepsis-Redox Bypass Hypothesis: Part B-Selenocompounds in the Management of Early Sepsis

Significance: Endothelial barrier damage, which is in part caused by excess production of reactive oxygen, halogen and nitrogen species (ROHNS), especially peroxynitrite (ONOO^-), is a major event in early sepsis and, with leukocyte hyperactivation, part of the generalized dysregulated immune response to infection, which may even become a complex maladaptive state. Selenoenzymes have major antioxidant functions. Their synthesis is related to the need to limit deleterious oxidant redox cycling by small selenocompounds, which may be of therapeutic cytotoxic interest. Plasma selenoprotein-P is crucial for selenium transport from the liver to the tissues and for antioxidant endothelial protection, especially against ONOO^-. Above micromolar concentrations, sodium selenite (Na₂SeO₃) becomes cytotoxic, with a lower cytotoxicity threshold in activated cells, which has led to cancer research. Recent Advances: Plasma selenium (<2% of total body selenium) is mainly contained in selenoprotein-P, and concentrations decrease rapidly in the early phase of sepsis, because of increased selenoprotein-P binding and downregulation of hepatic synthesis and excretion. At low concentrations, Na₂SeO₃ acts as a selenium donor, favoring selenoprotein-P synthesis in physiology, but probably not in the acute phase of sepsis. Critical Issues: The cytotoxic effects of Na₂SeO₃ against hyperactivated leukocytes, especially the most immature forms that liberate ROHNS, may be beneficial, but they may also be harmful for activated endothelial cells. Endothelial protection against ROHNS by selenoprotein-P may reduce Na₂SeO₃ toxicity, which is increased in sepsis. Future Direction: The combination of selenoprotein-P for endothelial protection and the cytotoxic effects of Na₂SeO₃ against hyperactivated leukocytes may be a promising intervention for early sepsis. Antioxid. Redox Signal. 37, 998-1029.

Nasopharyngeal lipidomic endotypes of infants with bronchiolitis and risk of childhood asthma: a multicentre prospective study

BACKGROUND

Bronchiolitis is the leading cause of hospitalisation of US infants and an important risk factor for childhood asthma. Recent evidence suggests that bronchiolitis is clinically heterogeneous. We sought to derive bronchiolitis endotypes by integrating clinical, virus and lipidomics data and to examine their relationship with subsequent asthma risk.

METHODS

This is a multicentre prospective cohort study of infants (age <12 months) hospitalised for bronchiolitis. We identified endotypes by applying clustering approaches to clinical, virus and nasopharyngeal airway lipidomic data measured at hospitalisation. We then determined their longitudinal association with the risk for developing asthma by age 6 years by fitting a mixed-effects logistic regression model. To account for multiple comparisons of the lipidomics data, we computed the false discovery rate (FDR). To understand the underlying biological mechanism of the endotypes, we also applied pathway analyses to the lipidomics data.

RESULTS

Of 917 infants with bronchiolitis (median age, 3 months), we identified clinically and biologically meaningful lipidomic endotypes: (A) cinicalclassiclipidmixed (n=263), (B) clinicalseverelipidsphingolipids-high (n=281), (C) clinicalmoderatelipidphospholipids-high (n=212) and (D) clinicalatopiclipidsphingolipids-low (n=161). Endotype A infants were characterised by 'classic' clinical presentation of bronchiolitis. Profile D infants were characterised by a higher proportion of parental asthma, IgE sensitisation and rhinovirus infection and low sphingolipids (eg, sphingomyelins, ceramides). Compared with endotype A, profile D infants had a significantly higher risk of asthma (22% vs 50%; unadjusted OR, 3.60; 95% CI 2.31 to 5.62; p<0.001). Additionally, endotype D had a significantly lower abundance of polyunsaturated fatty acids (eg, docosahexaenoic acid; FDR=0.01). The pathway analysis revealed that sphingolipid metabolism pathway was differentially expressed in endotype D (FDR=0.048).

CONCLUSIONS

In this multicentre prospective cohort study of infants with bronchiolitis, integrated clustering of clinical, virus and lipidomic data identified clinically and biologically distinct endotypes that have a significantly differential risk for developing asthma.Delete.

Data-driven clustering approach to identify novel phenotypes using multiple biomarkers in acute ischaemic stroke: A retrospective, multicentre cohort study

BACKGROUND

Acute ischaemic stroke (AIS) is a highly heterogeneous disorder and warrants further investigation to stratify patients with different outcomes and treatment responses. Using a large-scale stroke registry cohort, we applied data-driven approach to identify novel phenotypes based on multiple biomarkers.

METHODS

In a nationwide, prospective, 201-hospital registry study taking place in China between August 01, 2015 and March 31, 2018, the patients with AIS who were over 18 years of age and admitted to the hospital within 7 days from symptom onset were included. 92 biomarkers were included in the analysis. In the derivation cohort (n=9539), an unsupervised Gaussian mixture model was applied to categorize patients into distinct phenotypes. A classifier was developed using the most important biomarkers and was applied to categorize patients into their corresponding phenotypes in an validation cohort (n=2496). The differences in biological features, clinical outcomes, and treatment response were compared across the phenotypes.

FINDINGS

We identified four phenotypes with distinct characteristics in 9288 patients with non-cardioembolic ischaemic stroke. Phenotype 1 was associated with abnormal glucose and lipid metabolism. Phenotype 2 was characterized by inflammation and abnormal renal function. Phenotype 3 had the least laboratory abnormalities and small infarct lesions. Phenotype 4 was characterized by disturbance in homocysteine metabolism. Findings were replicated in the validation cohort. In comparison with phenotype 3, the risk of stroke recurrence (adjusted hazard ratio [aHR] 2.02, 95% confidence intervals [CI] 1.04-3.94), and mortality (aHR 18.14, 95%CI 6.62-49.71) at 3-month post-stroke were highest in phenotype 2, followed by phenotype 4 and phenotype 1, after adjustment for age, gender, smoking, drinking, history of stroke, hypertension, diabetes mellitus, dyslipidemia, and coronary heart disease. The Monte Carlo simulation showed that the patients with phenotype 2 could benefit from high-intensity statin therapy.

INTERPRETATION

A data-driven approach could aid in the identification of patients at a higher risk of adverse clinical outcomes following non-cardioembolic ischaemic stroke. These phenotypes, based on different pathophysiology, can suggest individualized treatment plans.

FUNDING

Beijing Natural Science Foundation (grant number Z200016), Beijing Municipal Committee of Science and Technology (grant number Z201100005620010), National Natural Science Foundation of China (grant number 82101360, 92046016, 82171270), Chinese Academy of Medical Sciences Innovation Fund for Medical Sciences (grant number 2019-I2M-5-029).