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

Extracorporeal therapies in sepsis: a comprehensive review of the Selective Cytopheretic Device, Polymyxin B and Seraph cartridges

Sepsis, a dysregulated host response to infection, is a leading cause of morbidity and mortality in critically ill patients, despite advancements in antimicrobial therapies. Recent innovations in extracorporeal blood purification therapies, such as the Selective Cytopheretic Device (SCD), Polymyxin B Hemoperfusion Cartridge (PMX-HP), and Seraph 100 Microbind Affinity Blood Filter (Seraph), have demonstrated promising potential as adjuncts to conventional therapies. The SCD targets activated white blood cells, while PMX-HP binds endotoxins in Gram-negative sepsis. The Seraph targets a broad range of pathogens, including viruses, bacteria and fungi. Evidence from several clinical trials and observational studies indicate that these therapies can improve organ function, and potentially improve survival in patients with sepsis. Despite the strong pathophysiological rationale for using these devices in sepsis, conclusive evidence of their effectiveness remains limited. Multicenter randomized controlled trials are currently underway with each of these devices to establish their role in improving patient outcomes. Further research is needed to establish optimal protocols for their initiation, duration, and integration into standard sepsis management.

Design of a quinoxalinone-based AIE probe for the detection of ROS in in vitro and in vivo sepsis models

Sepsis is one of the major causes of long-term mortality; the identification of potential biomarkers and developing specific and sensitive imaging and detection methods are crucial for timely diagnosis and progression evaluation. Reactive oxygen species (ROS) may serve as a potential detection and imaging marker for sepsis. Herein, we designed and synthesized a near-infrared quinoxalone framework-based aggregation-induced emission probe (QuinoNS NPs). We evaluated the selectivity, cytotoxicity, and detection and imaging ability in an in vitro LPS induced inflammatory model and an in vivo sepsis model. The probe can respond to ROS, causing a blue shift in the fluorescence emission wavelength. The probe can achieve real-time imaging and detection of ROS in LPS induced sepsis models both in vitro and in vivo with quick response and a superior duration time without significant toxicity. This study provides new strategies and theoretical basis for imaging and diagnosis of inflammatory diseases such as sepsis.

Using Artificial Intelligence to Identify Three Presenting Phenotypes of Chiari Type-1 Malformation and Syringomyelia

BACKGROUND AND OBJECTIVES

Chiari type-1 malformation (CM1) and syringomyelia (SM) are common related pediatric neurosurgical conditions with heterogeneous clinical and radiological presentations that offer challenges related to diagnosis and management. Artificial intelligence (AI) techniques have been used in other fields of medicine to identify different phenotypic clusters that guide clinical care. In this study, we use a novel, combined data-driven and clinician input feature selection process and AI clustering to differentiate presenting phenotypes of CM1 + SM.

METHODS

A total of 1340 patients with CM1 + SM in the Park Reeves Syringomyelia Research Consortium registry were split a priori into internal and external cohorts by site of enrollment. The internal cohort was used for feature selection and clustering. Features with high Laplacian scores were identified from preselected groups of clinically relevant variables. An expert clinician survey further identified features for inclusion that were not selected by the data-driven process.

RESULTS

The feature selection process identified 33 features (28 from the data-driven process and 5 from the clinician survey) from an initial pool of 582 variables that were incorporated into the final model. A K-modes clustering algorithm was used to identify an optimum of 3 clusters in the internal cohort. An identical process was performed independently in the external cohort with similar results. Cluster 1 was defined by older CM1 diagnosis age, small syringes, lower tonsil position, more headaches, and fewer other comorbidities. Cluster 2 was defined by younger CM1 diagnosis age, more bulbar symptoms and hydrocephalus, small syringes, more congenital medical issues, and more previous neurosurgical procedures. Cluster 3 was defined by largest syringes, highest prevalence of spine deformity, fewer headaches, less tonsillar ectopia, and more motor deficits.

CONCLUSION

This is the first study that uses an AI clustering algorithm combining a data-driven feature selection process with clinical expertise to identify different presenting phenotypes of CM1 + SM.

Molecular endotypes in sepsis: integration of multicohort transcriptomics based on RNA sequencing

BACKGROUND

The heterogeneity of host responses in sepsis has hindered efforts to develop targeted therapies for this large patient population. Although growing evidence has identified sepsis endotypes based on the microarray data, studies using RNA-seq data-which offers higher sensitivity and a broader dynamic range-remain limited. We hypothesized that integrating RNA-seq data from patients with sepsis would reveal molecular endotypes with distinct biological and clinical signatures.

METHODS

In this meta-analysis, we systematically searched for publicly available RNA-seq datasets of sepsis. Using identified datasets, we applied a consensus clustering algorithm to identify distinct endotypes. To characterize the biological differences between these endotypes, we performed gene-set enrichment analysis and immune cell deconvolution. Next, we investigated the association between these endotypes and mortality risks. We finally developed gene classifiers for endotype stratification and validated our endotype classification by applying these classifiers to an external cohort.

RESULTS

A total of 280 adults with sepsis from four datasets were included in this analysis. Using an unsupervised approach, we identified three distinct endotypes: coagulopathic (n = 83, 30%), inflammatory (n = 118, 42%), and adaptive endotype (n = 79, 28%). The coagulopathic endotype exhibited upregulated coagulation signaling, along with an increased monocyte and neutrophil composition, although the adaptive endotype demonstrated enhanced adaptive immune cell responses, marked by elevated T and B cell compositions. The inflammatory endotype was characterized by upregulated TNF-α/NF-κB signaling and IL-6/JAK/STAT3 pathways with an increased neutrophil composition. Patients with the coagulopathic endotype had a significantly higher risk of mortality than those with the adaptive endotype (30% vs. 16%, odds ratio 2.19, 95% confidence interval 1.04-4.78, p = 0.04). To enable the practical application of these findings, we developed endotype classification models and identified 14 gene classifiers. In a validation cohort of 123 patients, we consistently identified these three endotypes. Furthermore, the mortality risk pattern was reproduced, with the coagulopathic endotype showing greater mortality risk than the adaptive endotype (34% vs 18%, p = 0.10).

CONCLUSIONS

This multicohort RNA-seq meta-analysis identified three biologically and clinically distinct sepsis endotypes characterized by coagulopathic, adaptive, and inflammatory responses. This endotype-based approach to patient stratification may facilitate the development of more precise therapeutic strategies for sepsis.

Macrophage Notch1 drives septic cardiac dysfunction by impairing mitophagy and promoting NLRP3 activation

BACKGROUND

Sepsis is a life-threatening condition with limited therapeutic options, characterized as excessive systemic inflammation and multiple organ failure. Macrophages play critical roles in sepsis pathogenesis. Although numerous studies support the critical role of Notch signaling in most inflammatory diseases, the function of Notch1 signaling in macrophages activation and its underlying molecular mechanism during sepsis has not been fully elucidated.

METHODS

We evaluated Notch1 expression in a lipopolysaccharide (LPS)-induced model of septic cardiac dysfunction. Using macrophage-specific Notch1 knockout mice (NOTCH1ΔMyelo) in conjunction with AAV-F4/80-mediated NICD1 overexpression, we investigated the impact of Notch1 on septic cardiac injury. LPS-stimulated bone marrow-derived macrophages (BMDMs) were analyzed by flow cytometry and ELISA to assess mitochondrial damage and inflammasome activation. Mitophagy flux and related protein levels were quantified, and a mitophagy inhibitor was applied to further delineate Notch1's in vivo role. Downstream targets of Notch1 were identified and validated via ChIP-qPCR and luciferase reporter assays.

RESULTS

Intraperitoneal injection of LPS markedly impaired cardiac function, increased macrophage infiltration, and elevated Notch1 expression compared with PBS-treated controls. Notch1 expression was inversely correlated with cardiac performance in LPS-treated mice. Notably, macrophage-specific deletion of Notch1 significantly improved cardiac function, whereas NICD1 overexpression worsened LPS-induced cardiac injury. NOTCH1ΔMyelo macrophages showed reduced mitochondrial damage and diminished activation of NLRP3-dependent caspase-1. Moreover, LPS induced mitophagy, an effect that was further enhanced by Notch1 knockout. Mechanistically, ChIP-seq and qPCR analyses revealed that NICD1 upregulates Mst1 transcription. Furthermore, overexpression of Mst1 counteracted the increased mitophagy in Notch1-deficient macrophages, resulting in elevated mitochondrial reactive oxygen species production, inflammatory cytokine secretion, and caspase-1 activation during prolonged LPS stimulation.

CONCLUSION

Our study uncovers a novel role for Notch1 in exacerbating LPS-induced septic cardiac dysfunction by suppressing mitophagy in macrophages. These findings suggest that targeting Notch1 may offer a promising therapeutic strategy to mitigate sepsis-induced inflammation by restoring proper mitophagy.

Analysis of the heterogeneous treatment effect of vasoactive drug dosage and time on hospital mortality across different sepsis phenotypes: a retrospective cohort study

BACKGROUND

The heterogeneity of sepsis poses challenges for the individualized treatment of vasoactive drugs.

METHODS

This study used data from ICUs in MIMIC-IV (2008-2019) and eICU (2014-2015) databases, identified sepsis by sepsis-3 criteria, and stratified sepsis into phenotypes by consensus K-means. The norepinephrine equivalence (NEE) formula balance treatment of different vasoactive drugs, with NEE captured hourly for up to 72 h to record both time of use and dosage. The logistic regression model, including phenotype-dosage-time interactions, examined heterogeneous treatment effects on hospital mortality. To address confounding, three models were fitted: Model 1 unadjusted, Model 2 adjusted for age and sex, and Model 3 additionally included 7 clinical variables identified via machine learning and directed acyclic graph. Nonlinear dosage was further analyzed based on restricted cubic splines. P values and P for interaction were Bonferroni-adjusted.

RESULTS

A total of 54,673 sepsis patients were included for phenotype identification, and 8,803 patients were further analyzed to evaluate heterogeneous treatment effect of vasoactive drugs. Four sepsis phenotypes were identified: A, B, C and D. Phenotype D was the most severe subgroup, followed by phenotype C, while phenotypes A and B were mild subgroups. In Model 3, each 0.05 μg/kg/min increase in NEE dosage was linked to higher hospital mortality (OR 1.328, 95% CI 1.314-1.342; p < 0.001). Longer NEE time of use also significantly increased mortality risk (OR 1.006, 95% CI 1.005-1.007; p < 0.001). In addition, these associations varied significantly by phenotype (P for interaction < 0.001). In RCS model, phenotype A consistently showed higher mortality than the other phenotypes at NEE dosages of 0.1-0.5 µg/kg/min, with this gap increasing over time, showing a clear dosage-time dependence. Phenotype B displayed lower overall mortality but the steepest relative risk of hospital mortality increased as dosage and time (OR of dosage: 1.309; OR of time: 1.005) in Model 3. Phenotype C reached the highest mortality risk when dosages exceeded 0.5 µg/kg/min, which was dosage dependence. Finally, phenotype D followed a U-shaped curve in RCS model, and minimum mortality was around 20% at 0.03-0.05 µg/kg/min.

CONCLUSIONS

Sepsis phenotypes differ significantly in their treatment effects of vasoactive drug dosage and time of use, indicating the need for phenotype-specific treatment strategies to improve outcomes.

Endotoxin hemoadsorption in refractory septic shock with multiorgan dysfunction and extreme endotoxin activity

Endotoxin septic shock is marked by severe organ failure and mortality rate that exceeds fifty percent, underscoring the critical need to tailor management strategies. Monitoring -endotoxin activity can guide the initiation and direction of adjunctive treatment for refractory septic shock through hemoadsorption. Thus, intervening based on the pathophysiological foundation may potentially improve outcomes. This represents a step towards precision medicine in the management of septic shock adjunctive therapies, addressing a knowledge gap in this pathology that remains insufficiently defined. Despite its potential, in the setting of refractory septic shock and multiorgan dysfunction with extreme endotoxin activity (EAA ≥ 0.9), the data about efficacy of endotoxin hemoadsorption is scarce.

Identifying a unique signature of sepsis in patients with pre-existing cirrhosis

BACKGROUND

The pre-existing diagnosis of cirrhosis is a complicating factor in the progression and prognosis of sepsis; however, the unique epidemiology, sepsis characteristics, and underlying mechanisms of immune dysregulation in sepsis among patients with cirrhosis remain incompletely understood. Our primary objective was to identify clinical outcomes and biological characteristics that differ between patients with and without cirrhosis among critically ill patients with sepsis.

METHODS

We analyzed data from a prospective cohort of critically ill patients presenting to single center with sepsis. Subjects were followed for 6 days for the development of acute respiratory distress syndrome (ARDS) and acute kidney injury (AKI), and 30 days for mortality. Inflammatory, endothelial, and coagulopathic proteins were measured in plasma collected at ICU admission in a subset of patients. We determined associations of cirrhosis with outcomes using multivariable logistic regression adjusting for pre-specified confounders. We tested differences in plasma protein levels by cirrhosis diagnosis using the Wilcoxon Rank-sum test.

RESULTS

We enrolled 2962 subjects, 371 (13%) of whom had a pre-existing diagnosis of cirrhosis. Patients with cirrhosis had higher severity of illness scores, were more likely to have an abdominal source of sepsis, and had more significant clinically measured coagulation abnormalities relative to patients without cirrhosis. In multivariate analysis, cirrhosis was associated with higher AKI risk (adjusted OR 1.65; 95% CI 1.21 to 2.26; P = 0.002), and 30-day mortality (adjusted OR 1.38; 95% CI 1.05 to 1.82; P = 0.022). There was no significant difference in risk for ARDS (adjusted OR 1.02; 95% CI 0.69 to 1.50; P = 0.92). Cirrhosis was associated with higher plasma levels of angiopoietin-2 (P < 0.001), von Willebrand factor (P < 0.001), and soluble thrombomodulin (P < 0.001), as well as lower levels of interleukin (IL)-10 (P < 0.001), IL-1β (P = 0.008), and IL-1RA (P = 0.036). There were no significant differences in levels of IL-6 (P = 0.30).

CONCLUSIONS

We identified associations between pre-existing cirrhosis and endothelial injury, AKI, and mortality in sepsis. Patients with pre-existing cirrhosis who develop sepsis may display a unique phenotype of endothelial dysfunction that requires unique targeted approaches.

Biomarkers as Beacons: Illuminating Sepsis-Associated Hepato-Renal Injury

Sepsis, defined as a dysregulated host response to infection, is one of the leading causes of mortality worldwide. It unleashes in the organism a cascade of molecules, cytokines, and proteins which leads to an inflammatory storm. If this response to infection is uncontrolled, any organ is susceptible to damage. Acute kidney injury (AKI) is one of the most frequent organ dysfunctions in septic patients, and while it can be reversible, its presence leads to a higher burden of morbidity and mortality. While serum creatinine is essential in evaluating kidney function, the pathophysiology of AKI is not completely elucidated, and a plethora of novel biomarkers have been studied in the hope of an early diagnosis and fast treatment. While the liver is not as affected by sepsis, it plays an important role as a guardian by providing acute-phase proteins, activating neutrophils, and controlling iron balance. Acute liver failure (ALF) could impair the organism's capacity to contain and eliminate pathogens. Some molecules have been associated with either AKI or ALF, although biomarkers specific for organ dysfunction are difficult to validate. The aim of this review is to understand the role of several molecules in the pathophysiology of sepsis and their clinical ability for diagnosing or predicting sepsis-induced hepato-renal dysfunction.

Diagnostic challenges in pulmonary lymphomatous spread mimicking ARDS in an AIDS patient: a case report

BACKGROUND

Immunocompromised individuals, particularly those with AIDS, are at increased risk of developing lymphoproliferative tumours and opportunistic infections. Radiologic findings alone may not always distinguish between these entities.

CASE PRESENTATION

We describe the case of a patient with acquired immunodeficiency syndrome (AIDS) with rapidly worsening dyspnoea and clinical signs suggestive of acute respiratory distress syndrome (ARDS). Despite initial concerns for ARDS, autopsy revealed an advanced-stage, aggressive lymphoma as the underlying cause. This case highlights the challenge of differentiating ARDS from lymphoma in AIDS patients, especially when atypical radiologic findings, such as nodular opacities, are present.

CONCLUSIONS

The diagnosis of ARDS relies on imaging, oxygenation abnormalities, and clinical timing. However, various infectious and non-infectious conditions can mimic ARDS, making an accurate differential diagnosis essential. This case adds to the literature by underscoring the importance of considering lymphoproliferative disorders in AIDS patients presenting with respiratory distress, especially in the absence of typical lymphoma-related symptoms.

Maximum heart rate and mortality in sepsis patients: a retrospective cohort study

The effects of maximum heart rate (MHR) on the prognosis of sepsis patients are still being determined, and the most optimal MHR is yet to be recommended. We aimed to determine the association between MHR and 28-day mortality in sepsis patients. A total of 269 patients with sepsis were enrolled in this retrospective cohort study. The patients were all from a 28-bed multi-disciplinary ICU in a tertiary teaching hospital. Relevant information about sepsis patients was extracted according to specific inclusion and exclusion criteria. Examined parameters such as the MHR, blood pressure, laboratory variable, primary disease, and comorbidities were collected. The study outcome was a 28-day death. A Cox proportional hazard model and restricted cubic splines (RCS) were used to elucidate the association between MHR and the 28-day mortality of sepsis patients. Sensitivity analyses were also conducted on our results. Finally, we test and verify our results using a validation cohort from a large public database. In patients with sepsis, the 28-day survivors have the lower MHR than the non-survivors (106 vs. 118, P<0.001). After univariate and multivariate Cox regression analyses, the MHR was significantly correlated with 28-day mortality [hazard ratio 1.013, 95% confidence interval (CI) 1.004-1.022, P = 0.004; hazard ratio 1.013, 95% CI 1.004-1.021, P = 0.004]. The patients in the highest quartile of MHR had a higher risk for 28-day mortality than other patients based on the Kaplan-Meier survival curves (log-rank P<0.001). This correlation remained remarkable in subgroup analysis. A typical U-shaped curve relationship was observed between MHR and the 28-day mortality as a continuous variable using the RCS model. The curve shows that either too high or too low MHR will increase the risk of death. The sepsis patients who had a MHR ranging from approximately 70 to 110 beats per minute (bpm) had the lowest 28-day mortality rate. A similar trend was identified in the validation cohort. High or low MHR was associated with an elevated risk for 28-day mortality in patients with sepsis.

Nomogram for Predicting Sepsis After Percutaneous Transhepatic Cholangioscopic Lithotripsy

Purpose

Sepsis is a possible complication of percutaneous transhepatic cholangioscopic lithotripsy (PTCSL) for hepatolithiasis, but risk assessment tools are lacking. This study aimed to identify predictors of sepsis after PTCSL and develop a predictive nomogram.

Patients and Methods

In this nested case‒control study, the data from 298 patients who underwent 528 PTCSL sessions between 1 January 2016 and 1 July 2024 were retrospectively reviewed. All sessions demonstrating sepsis complications were included in the sepsis group. For each session in the sepsis group, two treatment date-matched sessions not demonstrating sepsis were randomly selected via a nested case‒control design. All the matched sessions were divided into training and validation sets. Least absolute shrinkage and selection operator (LASSO) analysis was conducted to preliminarily select predictors of sepsis complications. Multivariable logistic regression was performed to identify factors for constructing the nomogram.

Results

Sepsis was diagnosed in 46 patients (53 sessions), for an incidence of 10.69% (53 among 496 sessions). Three characteristic variables were included in the model: operation technique (odds ratio [OR]=0.170, 95% confidence interval [CI]: 0.048-0.599, P=0.006), cirrhosis (OR=3.769, 95% CI: 1.474-9.638, P=0.006), and postoperative prophylactic dexamethasone (OR=0.267, 95% CI: 0.101-0.703, P=0.008). The area under the curve (AUC) for the nomogram was 0.756 (95% CI, 0.658-0.853) in the training set and 0.762 (95% CI, 0.618-0.906) in the validation set, demonstrating relatively high discriminability. The calibration curves demonstrated the consistency between the predicted and actual values. Decision curve analysis indicated that the nomogram offers net clinical benefits.

Conclusion

The operation technique, cirrhosis, and postoperative prophylactic dexamethasone may predict the occurrence of sepsis after PTCSL. We developed a nomogram to predict sepsis complications following PTCSL and demonstrated its relatively strong performance.

Evaluating the obesity paradox in patients with sepsis and cancer

BACKGROUND

The obesity paradox, where higher body mass index (BMI) is associated with improved survival, is observed in sepsis and cancer individually. However, its effect in patients with both conditions is unclear. The objective of this study is to evaluate the obesity paradox in intensive care unit (ICU) patients with sepsis and cancer and examine whether BMI's impact on mortality varies across patient phenotypes.

METHODS

Data from the Medical Information Mart for Intensive Care IV database were analyzed. Patients were categorized by BMI into underweight (<18.5 kg/m2, n = 173), normal weight (18.5-24.9 kg/m2, n = 1283), overweight (25-29.9 kg/m2, n = 1498), and obesity (≥30 kg/m2, n = 960). The primary outcome was 28-day mortality, with secondary outcomes including 6-month mortality, 1-year mortality, length of ICU stay, continuous renal replacement therapy usage, and invasive ventilation usage. Multivariable logistic regression and restricted cubic splines were used to explore the BMI-mortality relationship, and unsupervised clustering was performed to identify patient phenotypes.

RESULTS

Among 3914 patients, obesity was associated with lower mortality. Clustering revealed four distinct phenotypes, with the protective effect of obesity being more evident in patients with lower Sequential Organ Failure Assessment (SOFA) scores (Cluster1, Cluster2 and Cluster3).

CONCLUSIONS

The obesity paradox is evident in both short-term outcome (28-day mortality) and long-term outcomes (6-month and 1-year mortality) among patients with sepsis and cancer, particularly in those presenting with lower disease severity. These findings highlight the need for personalized treatment approaches in this complex patient population.

Association between neutrophil-platelet ratio and 28-day mortality in patients with sepsis: a retrospective analysis based on MIMIC-IV database

BACKGROUND

The immune system and inflammation are intimately linked to the pathophysiology of sepsis. The neutrophil‒platelet ratio (NPR), associated with inflammation and immunology, may be useful in predicting sepsis outcomes. According to earlier research, the NPR is linked to the prognosis of several diseases. This study aimed to investigate the connection between the NPR and unfavorable outcomes in patients with sepsis.

METHODS

We retrieved patient clinical data from the Medical Information Mart for Intensive Care IV database (MIMIC-IV 2.2) based on the inclusion and exclusion criteria. The NPR quartile was used to divide the population into four groups. 28-day mortality was the main result, whereas 90-day mortality was the secondary result. The Cox regression model, Kaplan‒Meier survival curve, and limited cubic spline were used to examine the associations between the NPR and the negative outcomes of sepsis. Subgroup analysis was also conducted. At the same time, we used Latent Class Trajectory Model (LCTM) to assess the trajectory of NPR within six days of ICU admission, and to assess the relationship between NPR trajectory and mortality at 28 and 90 days.

RESULTS

This study included 3339 patients. Quartile 4 had the greatest 28-day and 90-day mortality rates, according to the Cox regression model and Kaplan‒Meier survival curve. A J-shaped relationship between the NPR and mortality was found in restricted cubic spline investigations. This means higher and lower NPRs were linked to higher mortality, with NPR = 3.81 as the tipping point. A total of 434 patients were included in the trajectory analysis, and three trajectory patterns were identified. Patients with sepsis had an increased mortality rate in the slow-decline group compared with the stable development group.

CONCLUSION

The NPR has prognostic value for patients with sepsis, and there is a J-shaped relationship between the two variables. Patients with sepsis who have a slowly declining NPR have an increased mortality rate.

CLINICAL TRIAL

Not applicable.

Comparison of different AI systems for diagnosing sepsis, septic shock, and cardiogenic shock: a retrospective study

Sepsis, septic shock, and cardiogenic shock are life-threatening conditions associated with high mortality rates, but differentiating them is complex because they share certain symptoms. Using the Medical Information Mart for Intensive Care (MIMIC)-III database and artificial intelligence (AI), we aimed to increase diagnostic precision, focusing on Bayesian network classifiers (BNCs) and comparing them with other AI methods. Data from 5970 adults, including 950 patients with cardiogenic shock, 1946 patients with septic shock, and 3074 patients with sepsis, were extracted for this study. Of the original 51 variables included in the data records, 12 were selected for constructing the predictive model. The data were divided into training and validation sets at an 80:20 ratio, and the performance of the BNCs was evaluated and compared with that of other AI models, such as the one rule classifier (OneR), classification and regression tree (CART), and an artificial neural network (ANN), in terms of accuracy, sensitivity, specificity, precision, and F1-score. The BNCs exhibited an accuracy of 87.6% to 91.5%. The CART model demonstrated a notable 91.6% accuracy when only three decision levels were used, whereas the intricate ANN model reached 90.5% accuracy. Both the BNCs and the CART model allowed clear interpretation of the predictions. BNCs have the potential to be valuable tools in diagnostic tasks, with an accuracy, sensitivity, and precision comparable, in some cases, to those of ANNs while demonstrating superior interpretability.

Detection and Evaluation of Procalcitonin Variants As Diagnostic Tools in Systemic Inflammation

BACKGROUND

Procalcitonin is an indicator of systemic inflammation associated with major surgery or sepsis. Procalcitonin exists in a full-length and truncated variant as a result of dipeptidylpeptidase-4 (DPP4)-cleavage. We recently identified differential biological activity of both variants. Here, we present an immunoassay-based method for the separate detection of procalcitonin variants and correlation to clinical data in patients with severe systemic inflammation.

METHODS

Rabbits were immunized with peptides of N-terminal sequences of both human procalcitonin variants and polyclonal antibodies purified from rabbit plasma. Antibodies were used for the detection of procalcitonin variants in an indirect sandwich enzyme-linked immunosorbent assay (ELISA) using a commercially available monoclonal anti-procalcitonin antibody as capture. Serum was collected from 19 septic patients exhibiting hyperprocalcitonemia as part of a cross-sectional study; clinical data were analyzed and correlated with procalcitonin variant measurements. DPP4 activity was determined by a DPP4 activity assay.

RESULTS

Purified antibodies allowed for the separate detection of both procalcitonin variants in all patients. Levels of truncated procalcitonin (truncPCT) correlated with DPP4-activity (Pearson's R = 0.85, P < .001) and negatively correlated with patients' Sequential Organ Failure Score (SOFA) scores (Pearson's R = -0.56, P = .013). In contrast, the correlation between full-length procalcitonin (fullPCT) and SOFA scores was positive (Pearson's R = 0.56, P = .013). Separation of the patient collective into groups with higher amounts of fullPCT versus truncPCT revealed higher SOFA scores in patients with fullPCT > truncPCT (mean ± standard error of the mean; 11. 3 ± 0.8 vs 6. 1 ± 1.5, P = .003). Patients with fullPCT > truncPCT showed a tendency towards higher doses of vasopressor (0. 2 ± 0.1 vs 0. 1 ± 0.03 µg/kg/min norepinephrine within the first 24 hours after sepsis diagnosis, P = .062) and exhibited higher creatinine (2. 0 ± 0.2 vs 1. 4 ± 0.3mg/dL, P = .019) and leukocyte levels (31. 0 ± 5.4 vs 12. 8 ± 1.9cells/µL, P = .012). In addition, patients with fullPCT > truncPCT were more often subjected to treatment with hydrocortisone (49.0 vs 0%, P = .018).

CONCLUSIONS

Polyclonal antibodies generated using procalcitonin N-terminal variant peptides as immunogens are suitable for procalcitonin variant assessment. The separate detection of procalcitonin variants may offer additional diagnostic value and can be correlated with organ dysfunction and clinical outcomes in patients with systemic inflammation.

Ig-M and Ig-A Enriched Ig-G Infusion as Adjuvant Therapy in the Critically ill Patients Experiencing SARS-CoV-2 Severe Infection

Introduction: SARS-CoV-2 in patients who need Intensive Care (ICU) is associated with a mortality rate ranging from 10 to 40%-45%, with an increase in morbidity and mortality in presence of sepsis. Methods: We assumed that immunoglobulin (Ig) M and IgA enriched IgG (IGAM) therapy may support SARS COV-2 sepsis-related phase improving patient outcome. We conducted a retrospective case-control study on all the patients admitted to our ICU during the three pandemic waves between February 2020 and April 2021. Upon ICU admission, patients received anticoagulants with the standard supportive treatment (ST) ± IGAM therapy. After matching for the baseline characteristics and treatments, the patients receiving IGAM therapy too (group A), were compared with those undergoing ST (group B) only. Results: 85 patients were enrolled in group A, whereas 111 in group B. The mortality resulted lower in group A [37.6% versus 55.8%, OR: 0.7 (02-08), P = .01)]. A logistic regression analysis identified IGAM treatment as a survival predictor [OR: 0.35 (95%CI, 0.2-0.8)], whereas experiencing a super-infection [OR: 1.88 (95%CI, 1.5-4.9)] and a septic shock [OR: 1.92 (95%CI, 1.4-4.3)] as predictors of death. On day 7, the probability of dying was 3 times higher in patients treated with ST only. Variable life adjustment display (VLAD) was equal to 2.4 in group A, while - 2.2 group B (in terms of lives saved in relation with those expected, in according with Simplified Acute Physiology Score II (SAPS II) score. Conclusion: The treatment based on IGAM infusion seems to give an advantage chance of survival in SARS-CoV-2 severe infection. Further prospective studies are warranted.

Sepsis-induced immunosuppression: mechanisms, biomarkers and immunotherapy

Sepsis, a life-threatening organ dysfunction resulting from a dysregulated host response to infection, initiates a complex immune response that varies over time, characterized by sustained excessive inflammation and immunosuppression. Sepsis-induced immunosuppression is now recognized as a major cause of septic death, and identifying effective strategies to counteract it poses a significant challenge. This immunosuppression results from the disruption of immune homeostasis, characterized by the abnormal death of immune effector cells, hyperproliferation of immune suppressor cells, release of anti-inflammatory cytokines, and expression of immune checkpoints. Preclinical studies targeting immunosuppression, particularly with immune checkpoint inhibitors, have shown promise in reversing immunocyte dysfunctions and establishing host resistance to pathogens. Here, our review highlights the mechanisms of sepsis-induced immunosuppression and current diagnostic biomarkers, as well as immune-enhancing strategies evaluated in septic patients and therapeutics under investigation.

Predictors of Cumulative 90-D Mortality for Septic Patients Undergoing Abdominal Surgery

INTRODUCTION

Septic surgical patients undergoing emergency general surgery represent a distinct population with unique challenges. This study aimed to identify predictors of cumulative 90-d mortality, including clinical and socioeconomic factors, and to analyze causes of death in this cohort.

METHODS

A retrospective analysis was conducted on patients admitted to a surgical intensive care unit from 2011 to 2019 with sepsis (sequential organ failure assessment score ≥2) undergoing emergency intra-abdominal surgery (n = 498). Demographics, comorbidities, sepsis presentation, and socioeconomic metrics, including the area deprivation index (ADI), were analyzed. Independent predictors of mortality were identified using multiple logistic regression. The causes of death were categorized and analyzed.

RESULTS

Among 498 patients, 46% (n = 229) died within 90 d. Nonsurvivors were older (65 ± 13.7 versus 61.2 ± 13.5 y, P < 0.01), more often transferred from external facilities (59% versus 46%, P < 0.01), and had higher rates of liver disease, chronic kidney disease, metastatic cancer, obesity, and higher Charlson comorbidity index scores (P < 0.01 for all). Independent predictors of 90-d mortality included admission sequential organ failure assessment scores, serum lactate, obesity, ADI, Charlson comorbidity index, and transfer status. ADI remained a significant predictor, while the distressed communities index did not. Of the deaths, 76.9% were in-hospital deaths, with intra-abdominal catastrophes (35.4%), multisystem organ failure (25.2%), and pulmonary causes (16.4%) as the most common causes.

CONCLUSIONS

Intra-abdominal catastrophes, multiorgan failure, and pulmonary complications are leading causes of death in septic emergency general surgery patients. ADI is a robust socioeconomic predictor of mortality, underscoring the need for integrating social determinants into risk assessment and tailored care strategies. Developing comprehensive risk models may enhance prognostication and guide clinical decision-making in this critical population.

Immunomodulation in Pediatric Sepsis: A Narrative Review

Pediatric sepsis presents a unique clinical challenge due to the distinct characteristics of the developing immune system. The immune response in children differs significantly from that in adults, exhibiting a unique combination of resistance, disease tolerance, and resilience. These factors influence the clinical presentation and prognosis of pediatric patients with sepsis. Over the past few years, various studies have explored the role of immunomodulatory therapies in managing sepsis, including the use of immunoglobulins, corticosteroids, monoclonal antibodies, and immunostimulatory treatments. However, the heterogeneity of the clinical presentations and individual responses makes it difficult to identify universally effective treatments. Recent research has highlighted the importance of a personalized approach based on specific biomarkers and patient phenotyping. Extracorporeal blood purification techniques have emerged as promising strategies for the modulation of hyperinflammation. However, strong evidence supporting their routine use in pediatric sepsis is lacking. This review provides a comprehensive overview of the current knowledge of the immune response in pediatric sepsis and discusses the main immunomodulatory strategies and future perspectives for personalized therapy. A deeper understanding of the immunological differences between children and adults could improve the prognosis and treatment efficacy, paving the way for new approaches to pediatric sepsis management.

Prognostic value of multivariate logistic regression analysis and amyloid A lactate monitoring in patients with severe pneumonia-associated sepsis

BACKGROUND

Patients with severe pneumonia-associated sepsis often face high mortality rates, highlighting the need for simple and effective prognostic biomarkers. This study aimed to investigate the prognostic significance of serum amyloid A (SAA) and blood lactate (Lac) levels using multivariate logistic regression.

METHOD

This was a retrospective study conducted from January 2021 to December 2023, which included 156 patients diagnosed with severe pneumonia. Of these, 54 developed sepsis (septic group) while 102 did not (non-septic group). Clinical data, SAA, and Lac levels were compared between the groups. Multivariate logistic regression was employed to identify factors influencing the onset of severe pneumonia-associated sepsis and to assess the prognostic significance of SAA and Lac.

RESULT

Significant differences were found in APACHE II score, SOFA score, age, mechanical ventilation, SAA, and Lac levels between the septic and non-septic groups (P < 0.05). Logistic regression analysis identified age, SOFA score, APACHE II score, mechanical ventilation, SAA, and Lac as influencing factors for severe pneumonia-associated sepsis (P < 0.05). Patients with poor prognosis (PP) had significantly elevated SAA and Lac levels compared to those with good prognosis (GP) (P < 0.05). Among septic patients, those with PP had significantly higher SAA and Lac levels compared to those with GP (P < 0.05). Multivariate logistic regression revealed that advanced age, septic shock, elevated SAA levels, and increased Lac levels were predictors of PP (P < 0.05). The prognostic value of SAA and Lac was demonstrated by AUCs of 0.764 and 0.771, respectively. When combined, the AUC increased to 0.903 with a specificity of 95.00% and sensitivity of 80.25%.

CONCLUSION

Severe pneumonia-associated sepsis is influenced by age, SOFA score, APACHE II score, mechanical ventilation, SAA, and Lac levels. Elevated SAA and Lac levels are associated with PP and can provide prognostic information for adverse outcomes. While SAA and Lac show potential as biomarkers for predicting the prognosis of severe pneumonia-associated sepsis, their clinical utility should be considered in conjunction with other diagnostic and clinical factors for effective patient management and risk stratification.

CXCR4+ PD-L1+ neutrophils are increased in non-survived septic mice

The dysregulated host response to infections can lead to sepsis, a complex disease characterized by a spectrum of clinical phenotypes. Using scRNA-seq, we analyzed the immune cell of survived and non-survived CLP-septic mice to gain insights into the immunological mechanisms by which neutrophils contribute to the hyperinflammatory phenotype. Our findings reveal that non-survived mice exhibit increased frequencies of immature CXCR4+ PD-L1+ neutrophils in the bloodstream, accompanied by an accumulation of trafficking-specific CXCR4+ PD-L1+ neutrophils into the lungs. The IFN-gamma and LPS promote the PD-L1 expression on neutrophils and an activation profile associated with inflammation and organ damage. Notably, abrogating the IFN-gamma reduced susceptibility to CLP-sepsis and diminished CXCR4+ PD-L1+ neutrophils frequency. This study provides insights into the immune cell activation profiles associated with the worsening of the CLP-sepsis, and the CXCR4+ PD-L1+ neutrophils population highlighted here represents a promising target for therapeutic modulation in clinical sepsis hyperinflammatory phenotype.

Distinct immunological signatures define three sepsis recovery trajectories: a multi-cohort machine learning study

Importance

Understanding heterogeneous recovery patterns in sepsis is crucial for personalizing treatment strategies and improving outcomes.

Objective

To identify distinct recovery trajectories in sepsis and develop a prediction model using early clinical and immunological markers.

Design setting and participants

Retrospective cohort study using data from 28,745 adult patients admitted to 12 intensive care units (ICUs) with sepsis between January 2014 and December 2024.

Main outcomes and measures

Primary outcome was the 28-day trajectory of Sequential Organ Failure Assessment (SOFA) scores. Secondary outcomes included 90-day mortality and hospital length of stay.

Results

Among 24,450 eligible patients (mean [SD] age, 64.5 [15.3] years; 54.2% male), three distinct recovery trajectories were identified: rapid recovery (42.3%), slow recovery (35.8%), and deterioration (21.9%). The machine learning model achieved an AUROC of 0.85 (95% CI, 0.83-0.87) for trajectory prediction. Key predictors included initial SOFA score, lactate levels, and inflammatory markers. Mortality rates were 12.3, 28.7, and 45.6% for rapid, slow, and deterioration groups, respectively.

Conclusions and relevance

Early prediction of sepsis recovery trajectories is feasible and may facilitate personalized treatment strategies. The developed model could assist clinical decision-making and resource allocation in critical care settings.

Treatment effect of intravenous high-dose selenium in sepsis phenotypes: a retrospective analysis of a large multicenter randomized controlled trial

BACKGROUND

Treatment effect of high-dose intravenous selenium remains controversial in patients with sepsis or septic shock. Here, we reanalyzed data from the randomized placebo-controlled trial of Sodium Selenite and Procalcitonin Guided Antimicrobial Therapy in Severe Sepsis (SISPCT) to reveal possible treatment differences according to established sepsis phenotypes.

METHODS

In this secondary data analysis of the SISPCT trial all 1089 patients of the original study were included. Patients were assigned to one of the four phenotypes by comparing patient variables with the Sepsis Endotyping in Emergency Care (SENECA) validation cohort. Survival analyses were performed using Kaplan-Meier and log-rank tests.

RESULTS

No robust effect of selenium on mortality and other outcome parameters could be determined in any sepsis phenotype. Phenotype frequencies were markedly different in our study cohort compared to previous reports (α: 2.2%, β: 6.3%, γ: 68.0%, δ: 23.4%). Differences in mortality between the respective phenotypes were not significant overall; however, 28-day mortality showed a lower mortality for the α- (20.8%) and β-phenotype (20.3%), followed by the γ- (27.1%), and δ-phenotype (28.5%).

CONCLUSIONS

Application of the four sepsis phenotypes to the SISPCT study cohort showed discrete but non-significant mortality differences within 28 days. However, beneficial treatment effects of high-dose intravenous selenium were still not detectable after categorizing the SISPCT study cohort according to four phenotype criteria.

A narrative review of the pathophysiology of sepsis in sub-Saharan Africa: Exploring the potential for corticosteroid therapy

Sepsis remains a significant global health threat with a disproportionate burden in low-income countries including those in sub-Saharan Africa where case fatality rates are as high as 30% to 50%. Defined as a severe systemic response to infection, sepsis leads to widespread immune dysregulation and organ dysfunction, including adrenal insufficiency. Critical illness-related corticosteroid insufficiency (CIRCI) arises from dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis, altered cortisol metabolism, and tissue resistance to glucocorticoids, all of which can occur during sepsis. Clinical trials of corticosteroids for the treatment of patients with sepsis and septic shock have shown improvements in shock reversal, and in some studies, patient survival; however, their role in the treatment of sepsis in sub-Saharan Africa is unknown. The incidence of sepsis in sub-Saharan Africa is compounded by high rates of human immunodeficiency virus (HIV) and co-infections, including tuberculosis (TB), which is the leading cause of sepsis. Both HIV and TB can cause immune dysregulation and adrenal insufficiency, which may exacerbate CIRCI and prolong shock. Existing sepsis research has been predominantly conducted in high-income countries and has largely excluded people living with HIV or TB. Therefore, there is a need to better understand sepsis and CIRCI pathophysiology in the context of specific regional host and pathogen characteristics. In this narrative review, we explored the pathophysiology of sepsis in sub-Saharan Africa including the existing literature on the immune response to sepsis and the prevalence of adrenal insufficiency in patients with HIV and TB, with a focus on the implications for corticosteroid management. We found a compelling need to further evaluate corticosteroids for the treatment of sepsis in Africa.

Deciphering the impact of sepsis phenotypes on improving clinical outcome predictions: a multicenter retrospective analysis based on critical care in China

Sepsis is a clinically heterogeneous disease with high mortality. It is crucial to develop relevant therapeutic strategies for different sepsis phenotypes, but the impact of phenotypes on patients' clinical outcomes is unclear. This study aimed to identify potential sepsis phenotypes using readily available clinical parameters and assess their predictive value for 28-day clinical outcomes by logistic regression analysis. In this retrospective analysis, researchers extracted clinical data from adult patients admitted to the First Affiliated Hospital of Anhui Medical University between April and August 2022 and from the 2014-2015 eICU Collaborative Study database. K-Means clustering was utilized to identify and refine sepsis phenotypes, and their predictive performance was subsequently evaluated. Logistic regression models were trained independently for each phenotype and five-fold cross-validation was used to predict clinical outcomes. Predictive accuracy was then compared to traditional non-clustered prediction methods using model assessment scores. The study cohort consisted of 250 patients from the First Affiliated Hospital of Anhui Medical University, allocated in a 7:3 ratio for training and testing, respectively, and an external validation cohort of 3100 patients from the eICU Cooperative Research Database. The results of the phenotype-based prediction model demonstrated an improvement in F1 score from 0.74 to 0.82 and AUC from 0.74(95%CI 0.71-0.80) to 0.84(95%CI 0.82-0.87), and these results also highlight the superiority of clinical outcome prediction with the help of sepsis phenotypes over traditional prediction methods. Phenotype-based prediction of 28-day clinical outcomes in sepsis demonstrated significant advantages over traditional models, highlighting the impact of phenotype-driven modeling on clinical outcomes in sepsis.

A methodological systematic review of validation and performance of sepsis real-time prediction models

Sepsis real-time prediction models (SRPMs) provide timely alerts and may improve patient outcomes but face limited clinical adoption due to inconsistent validation methods and potential biases. Comprehensive evaluation, including external full-window validation with model- and outcome-level metrics, is crucial for real-world effectiveness, yet performance evidence remains scarce. This study systematically reviewed SRPM performance across validation methods, analyzing 91 studies from multiple databases. Only 54.9% applied full-window validation with both metric types. Performance decreased under external and full-window validation, with median AUROCs of 0.886 and 0.861 at 6- and 12-hours pre-onset, dropping to 0.783 in full-window external validation. Median Utility Scores declined from 0.381 in internal to -0.164 in external validation. Combining AUROC and Utility Score identified top-performing SRPMs in 18.7% of studies. Hand-crafted features significantly improved performance. Future research should focus on multi-center datasets, hand-crafted features, multi-metric full-window validation, and prospective trials to support clinical implementation.

Machine learning in point-of-care testing: innovations, challenges, and opportunities

The landscape of diagnostic testing is undergoing a significant transformation, driven by the integration of artificial intelligence (AI) and machine learning (ML) into decentralized, rapid, and accessible sensor platforms for point-of-care testing (POCT). The COVID-19 pandemic has accelerated the shift from centralized laboratory testing but also catalyzed the development of next-generation POCT platforms that leverage ML to enhance the accuracy, sensitivity, and overall efficiency of point-of-care sensors. This Perspective explores how ML is being embedded into various POCT modalities, including lateral flow assays, vertical flow assays, nucleic acid amplification tests, and imaging-based sensors, illustrating their impact through different applications. We also discuss several challenges, such as regulatory hurdles, reliability, and privacy concerns, that must be overcome for the widespread adoption of ML-enhanced POCT in clinical settings and provide a comprehensive overview of the current state of ML-driven POCT technologies, highlighting their potential impact in the future of healthcare.

The Effect of Severe Sepsis and Septic Shock Management Bundle (SEP-1) Compliance and Implementation on Mortality Among Patients With Sepsis : A Systematic Review

BACKGROUND

The Centers for Medicare & Medicaid Services (CMS) Severe Sepsis and Septic Shock Management Bundle (SEP-1) is now included in the Hospital Value-Based Purchasing (VBP) Program.

PURPOSE

To assess the evidence supporting SEP-1 compliance or SEP-1 implementation in improving sepsis mortality.

DATA SOURCES

PubMed, Web of Science, EMBASE, CINAHL Complete, and Cochrane Library from inception to 26 November 2024.

STUDY SELECTION

Studies of adults with sepsis that included 3- or 6-hour sepsis bundles defined by SEP-1 specifications.

DATA EXTRACTION

Article screening, full-text review, data extraction, and risk-of-bias assessment were independently performed by 2 authors. Level of evidence was determined using GRADE (Grading of Recommendations Assessment, Development and Evaluation) criteria and National Quality Forum criteria.

DATA SYNTHESIS

A total of 4403 unique references were screened, and 17 studies were included. Twelve studies assessed the relationship between SEP-1 compliance and mortality; 5 showed statistically significant benefit, whereas 7 did not. Among studies showing benefit, 1 did not adjust for confounders, 1 found benefit only among patients with severe sepsis, 1 included only patients with septic shock, and 1 included only Medicare beneficiaries. Five studies assessed the relationship between SEP-1 implementation and sepsis mortality; only 1 showed significant benefit, but it did not adjust for mortality trends before SEP-1 implementation. All 17 studies were observational, and none had low risk of bias.

LIMITATIONS

The conclusions are limited by the underlying quality of the available studies, as all were observational. Because there was considerable methodologic heterogeneity among the included studies, a meta-analysis was not performed as the results could have been misleading.

CONCLUSION

This review found no moderate- or high-level evidence to support that compliance with or implementation of SEP-1 was associated with sepsis mortality. CMS should reconsider the addition of SEP-1 to the Hospital VBP Program.

PRIMARY FUNDING SOURCE

None. (PROSPERO: CRD42023482787).

Emergency medicine updates: Management of sepsis and septic shock

INTRODUCTION

Sepsis is a common condition associated with significant morbidity and mortality. Emergency physicians play a key role in the diagnosis and management of this condition.

OBJECTIVE

This paper evaluates key evidence-based updates concerning the management of sepsis and septic shock for the emergency clinician.

DISCUSSION

Sepsis is a life-threatening syndrome, and rapid diagnosis and management are essential. Antimicrobials should be administered as soon as possible, as delays are associated with increased mortality. Resuscitation targets include mean arterial pressure ≥ 65 mmHg, mental status, capillary refill time, lactate, and urine output. Intravenous fluid resuscitation plays an integral role in those who are fluid responsive. Balanced crystalloids and normal saline are both reasonable options for resuscitation. Early vasopressors should be initiated in those who are not fluid-responsive. Norepinephrine is the recommended first-line vasopressor, and if hypotension persists, vasopressin should be considered, followed by epinephrine. Administration of vasopressors through a peripheral 20-gauge or larger intravenous line is safe and effective. Steroids such as hydrocortisone and fludrocortisone should be considered in those with refractory septic shock.

CONCLUSION

An understanding of the recent updates in the literature concerning sepsis and septic shock can assist emergency clinicians and improve the care of these patients.

Sepsis subphenotypes, theragnostics and personalized sepsis care

Heterogeneity between critically ill patients with sepsis is a major barrier to the discovery of effective therapies. The use of machine learning techniques, coupled with improved understanding of sepsis biology, has led to the identification of patient subphenotypes. This exciting development may help overcome the problem of patient heterogeneity and lead to the identification of patient subgroups with treatable traits. Re-analyses of completed clinical trials have demonstrated that patients with different subphenotypes may respond differently to treatments. This suggests that future clinical trials that take a precision medicine approach will have a higher likelihood of identifying effective therapeutics for patients based on their subphenotype. In this review, we describe the emerging subphenotypes identified in the critically ill and outline the promising immune modulation therapies which could have a beneficial treatment effect within some of these subphenotypes. Furthermore, we will also highlight how bringing subphenotype identification to the bedside could enable a new generation of precision-medicine clinical trials.

Preparing for future pandemics: Automated intensive care electronic health record data extraction to accelerate clinical insights

Background

Manual data abstraction from electronic health records (EHRs) for research on intensive care patients is time-intensive and challenging, especially during high-pressure periods such as pandemics. Automated data extraction is a potential alternative but may raise quality concerns. This study assessed the feasibility and credibility of automated data extraction during the coronavirus disease 2019 (COVID-19) pandemic.

Methods

We retrieved routinely collected data from the COVID-Predict Dutch Data Warehouse, a multicenter database containing the following data on intensive care patients with COVID-19: demographic, medication, laboratory results, and data from monitoring and life support devices. These data were sourced from EHRs using automated data extraction. We used these data to determine indices of wasted ventilation and their prognostic value and compared our findings to a previously published original study that relied on manual data abstraction largely from the same hospitals.

Results

Using automatically extracted data, we replicated the original study. Among 1515 patients intubated for over 2 days, Harris-Benedict (HB) estimates of dead space fraction increased over time and were higher in non-survivors at each time point: at the start of ventilation (0.70±0.13 vs. 0.67±0.15, P <0.001), day 1 (0.74±0.10 vs. 0.71±0.11, P<0.001), day 2 (0.77±0.09 vs. 0.73±0.11, P<0.001), and day 3 (0.78±0.09 vs. 0.74±0.10, P<0.001). Patients with HB dead space fraction above the median had an increased mortality rate of 13.5%, compared to 10.1% in those with values below the median (P<0.005). Ventilatory ratio showed similar trends, with mortality increasing from 10.8% to 12.9% (P=0.040). Conversely, the end-tidal-to-arterial partial pressure of carbon dioxide (PaCO₂) ratio was inversely related to mortality, with a lower 28-day mortality in the higher than median group (8.5% vs. 15.1%, P<0.001). After adjusting for base risk, impaired ventilation markers showed no significant association with 28-day mortality.

Conclusion

Manual data abstraction from EHRs may be unnecessary for reliable research on intensive care patients, highlighting the feasibility and credibility of automated data extraction as a trustworthy and scalable solution to accelerate clinical insights, especially during future pandemics.

Subphenotypes and the De Ritis ratio for mortality risk stratification in sepsis-associated acute liver injury: a retrospective cohort study

Background

Sepsis-associated liver injury (SALI) is associated with poor outcomes and increased mortality. However, effectively stratifying SALI patients according to prognosis remains challenging. This study evaluates laboratory-based clustering filters for stratifying SALI patients by 30-day mortality risk, utilizing data mining techniques for novel pattern discovery.

Methods

This retrospective cohort study analyzed SALI patients from two ICU databases: Medical Information Mart for Intensive Care (MIMIC)-IV database (n = 73,181, study period: 2008 to 2019) and Amsterdam UMC (n = 16,194, study period: 2003 to 2016). Patients were identified using Sepsis-3 criteria and liver injury markers. Risk stratification employed three laboratory-based approaches: (I) De Ritis ratio (aspartate aminotransferase/alanine aminotransferase), (II) R-factor (alanine aminotransferase and alkaline phosphatase relative to their upper limits of normal), and (III) alanine aminotransferase elevation. Kaplan-Meier analysis and multivariable Cox regression assessed the association between stratification methods and 30-day mortality risk.

Findings

SALI patients had almost a 2-fold higher risk of 30-day mortality than those without SALI (hazard ratio: 1.73; 95%-CI: 1.58-1.90, p < 0.0001). Each stratification method (I-III) successfully classified patients into statistically distinct risk strata. The De Ritis ratio emerged as the strongest prognostic differentiation method: a ratio ≤1 indicated no significant increase in mortality risk (hazard ratio: 0.86; 95%-CI: 0.68-1.09, p = 0.21), whereas ratios of 1-2 and ≥2 were significantly associated with higher mortality (hazard ratio: 1.56; 95%-CI: 1.37-1.78, p < 0.0001 and hazard ratio: 2.46; 95%-CI: 2.18-2.77, p < 0.0001, respectively). All findings were confirmed in the validation cohort.

Interpretation

The De Ritis ratio serves as a valuable prognostic tool for 30-day mortality in SALI patients. Our findings indicate that patients with a ratio ≥1 face significantly worse outcomes, highlighting the need for targeted interventions. These results refine risk stratification in SALI subphenotypes, enhancing our understanding of its prognostic implications.

Funding

This study received no external funding and was solely financed through the departmental resources of the authors.

Embracing a Penta helix hub framework for co-creating sustaining and potentially disruptive sterilization innovation that enables artificial intelligence and sustainability: A scoping review

The supply of safe pipeline medical devices is of paramount importance. Opportunities exist to transform reusable medical devices for improved processing that meets diverse patient needs. There is increased interest in multi-actor hub frameworks to meet innovation challenges globally. The purpose of this scoping paper was to identify critical decontamination and sterilization needs for the medtech and pharmaceutical sectors with a focus on understanding how to effectively use the Penta helix hub framework that combines academia, industry, healthcare, policy-makers/regulators and patients/society. A PRISMA scoping review of PubMed publications was conducted over the period 2010 to January 2025. Thirty of the 124 'helix hub' papers addressed innovation where only 3 of 16 healthcare-focused helices used or mentioned the need for key performance indicators (KPIs). Early-phase helix innovation ecosystems are mainly supported by qualitative or non-empirical data. This review explores multi-actor needs along with describing quantifiable KPIs at micro (end-user), meso (innovation hub) and macro (regional, national and international) levels. This integrated Penta hub approach will help to effectively plan, co-create, manage, analyse and utilize voluminous data, for example there are ca. 60,000 and 56,000 publications per year on artificial intelligence (AI) and medical devices respectively along, with some 35,000 adverse reports on devices submitted to the US FDA. This review addresses sustaining and potentially disruptive opportunities for decontamination and sterilization that includes the use of AI-enabled devices, bespoke training and sustainability.

Persistent mortality in critical COVID-19 ICU cases from wild-type to delta variant: A descriptive cohort study

The SARS-CoV-2 pandemic led to significant advancements in treatment and vaccination, contributing to a decline in overall COVID-19-related mortality. However, it remains unclear whether the mortality rate for critical cases treated on intensive care units has also decreased. This multicentric, multinational retrospective observational study analyzed 447 critically ill COVID-19 patients treated on ICUs across ten study centers in Germany and Austria. Patients were categorized into two periods: period 1 (March 2020 to May 2021, n = 316) and period 2 (June 2021 to January 2022, n = 131). Despite evolving treatment strategies and widespread vaccine availability in period 2, 30-day mortality remained unchanged (30% in period 1 vs. 37% in period 2; HR 1.26, 95% CI: 0.90-1.79, p = 0.181). Further outcomes, including ICU-free days (p = 0.735), ventilatory support-free days (p = 0.699), vasopressor-free days (p = 0.379), and dialysis-free days (p = 0.396), also showed no significant differences. Notably, 81% (106 of 131) of ICU patients in period 2 were unvaccinated, underscoring the persistent vulnerability of this group. These findings suggest that while public health measures reduced overall COVID-19 severity, critical illness remained highly lethal. Further research is needed to explore targeted interventions for high-risk ICU patients and to better understand the factors contributing to persistent mortality despite medical advancements.

Unsupervised clustering for sepsis identification in large-scale patient data: a model development and validation study

BACKGROUND

Sepsis is a major global health problem. However, it lacks a true reference standard for case identification, complicating epidemiologic surveillance. Consensus definitions have changed multiple times, clinicians struggle to identify sepsis at the bedside, and differing identification algorithms generate wide variation in incidence rates. The two current identification approaches use codes from administrative data, or electronic health record (EHR)-based algorithms such as the Center for Disease Control Adult Sepsis Event (ASE); both have limitations. Here our primary purpose is to report initial steps in developing a novel approach to identifying sepsis using unsupervised clustering methods. Secondarily, we report preliminary analysis of resulting clusters, using identification by ASE criteria as a familiar comparator.

METHODS

This retrospective cohort study used hospital administrative and EHR data on adults admitted to intensive care units (ICUs) at five Canadian medical centres (2015-2017), with split development and validation cohorts. After preprocessing 592 variables (demographics, encounter characteristics, diagnoses, medications, laboratory tests, and clinical management) and applying data reduction, we presented 55 principal components to eight different clustering algorithms. An automated elbow method determined the optimal number of clusters, and the optimal algorithm was selected based on clustering metrics for consistency, separation, distribution and stability. Cluster membership in the validation cohort was assigned using an XGBoost model trained to predict cluster membership in the development cohort. For cluster analysis, we prospectively subdivided clusters by their fractions meeting ASE criteria (≥ 50% ASE-majority clusters vs. ASE-minority clusters), and compared their characteristics.

RESULTS

There were 3660 patients in the development cohort and 3012 in the validation cohort, of which 21.5% (development) and 19.1% (validation) were ASE (+). The Robust and Sparse K-means Clustering (RSKC) method performed best. In the development cohort, it identified 48 clusters of hospitalizations; 11 ASE-majority clusters contained 22.4% of all patients but 77.8% of all ASE (+) patients. 34.9% of the 209 ASE (-) patients in the ASE-majority clusters met more liberal ASE criteria for sepsis. Findings were consistent in the validation cohort.

CONCLUSIONS

Unsupervised clustering applied to diverse, large-scale medical data offers a promising approach to the identification of sepsis phenotypes for epidemiological surveillance.

The Effect of Sepsis Recognition on Telemedicine Use in Rural Emergency Department Sepsis Treatment

Background: Provider-to-provider emergency department telehealth (tele-ED) has been proposed to improve rural sepsis care. The objective of this study was to measure the association between sepsis documentation and tele-ED use, treatment guideline adherence, and mortality. Methods: This analysis was a multicenter (n = 23) cohort study of sepsis patients treated in rural emergency departments (EDs) that participated in a tele-ED network between August 2016 and June 2019. The primary outcome was whether sepsis was documented explicitly in the clinical note impression in the local ED, and the primary exposure was rural tele-ED use, with secondary outcomes of time to tele-ED use, 3-h guideline adherence, and in-hospital mortality. Results: Data from 1,146 rural sepsis patients were included, 315 (27%) had tele-ED used and 415 (36%) had sepsis recognized in the rural ED. Tele-ED use was not independently associated with sepsis recognition (adjusted odds ratio [aOR]: 1.23, 95% confidence interval [CI]: 0.90-1.67). Sepsis recognition was associated with earlier tele-ED activation (adjusted hazard ratio 1.66, 95% CI: 1.28-2.15) and greater 3-h guideline adherence (aOR 1.37, 95% CI 1.03-1.83) Sepsis recognition was not independently associated with mortality (aOR 1.32, 95% CI 0.97-1.80). Conclusions: Although tele-ED care is a promising strategy to improve sepsis outcomes, its use was limited by under-recognition of sepsis in rural EDs.

Hemoperfusion Using the Oxiris Membrane in Septic Shock Patients with Preserved Kidney Function: A Case Series

Background/Objectives: Sepsis, a life-threatening condition caused by a dysregulated immune response to infection, is associated with high mortality. Endotoxin and cytokine overload play a crucial role in sepsis-induced organ dysfunction. The Oxiris® membrane, traditionally used as a hemofilter for renal replacement therapy, has demonstrated the capacity to adsorb endotoxins and cytokines. This study investigates the clinical effect during hemoperfusion with the Oxiris® membrane in patients with septic shock and preserved renal function. Methods: We present three adult patients with septic shock who were admitted to the intensive care unit with high vasopressor requirements and elevated inflammatory markers. As they were refractory to standard therapy and renal function was preserved, a 12-hour hemoperfusion session with an Oxiris® membrane was initiated. Hemodynamic parameters, inflammatory biomarkers, and endotoxin concentrations were evaluated before, during, and after hemoperfusion treatment. Results: All patients demonstrated hemodynamic stabilization, with norepinephrine support reduced by 10.3% to 70.0%. Key inflammatory markers decreased significantly, including interleukin-6 (-41.6% to -94.0%), procalcitonin (-29.3% to -49.5%), and C-reactive protein (4.7% to -37.2%). Endotoxin concentrations decreased by 62.0% and 13.6% in two of the three patients. No adverse effects related to hemoperfusion were observed. Conclusions: Hemoperfusion with the Oxiris® membrane effectively reduced vasopressor support, inflammatory markers, and endotoxin concentrations in patients with refractory septic shock. This approach may offer a novel strategy for early immune modulation in sepsis before renal dysfunction occurs. Further studies with larger cohorts are required to validate these findings and determine optimal treatment protocols.

Optimal Vasopressin Initiation in Septic Shock: The OVISS Reinforcement Learning Study

Importance

Norepinephrine is the first-line vasopressor for patients with septic shock. When and whether a second agent, such as vasopressin, should be added is unknown.

Objective

To derive and validate a reinforcement learning model to determine the optimal initiation rule for vasopressin in adult, critically ill patients receiving norepinephrine for septic shock.

Design, Setting, and Participants

Reinforcement learning was used to generate the optimal rule for vasopressin initiation to improve short-term and hospital outcomes, using electronic health record data from 3608 patients who met the Sepsis-3 shock criteria at 5 California hospitals from 2012 to 2023. The rule was evaluated in 628 patients from the California dataset and 3 external datasets comprising 10 217 patients from 227 US hospitals, using weighted importance sampling and pooled logistic regression with inverse probability weighting.

Exposures

Clinical, laboratory, and treatment variables grouped hourly for 120 hours in the electronic health record.

Main Outcome and Measure

The primary outcome was in-hospital mortality.

Results

The derivation cohort (n = 3608) included 2075 men (57%) and had a median (IQR) age of 63 (56-70) years and Sequential Organ Failure Assessment (SOFA) score at shock onset of 5 (3-7 [range, 0-24, with higher scores associated with greater mortality]). The validation cohorts (n = 10 217) were 56% male (n = 5743) with a median (IQR) age of 67 (57-75) years and a SOFA score of 6 (4-9). In validation data, the model suggested vasopressin initiation in more patients (87% vs 31%), earlier relative to shock onset (median [IQR], 4 [1-8] vs 5 [1-14] hours), and at lower norepinephrine doses (median [IQR], 0.20 [0.08-0.45] vs 0.37 [0.17-0.69] µg/kg/min) compared with clinicians' actions. The rule was associated with a larger expected reward in validation data compared with clinician actions (weighted importance sampling difference, 31 [95% CI, 15-52]). The adjusted odds of hospital mortality were lower if vasopressin initiation was similar to the rule compared with different (odds ratio, 0.81 [95% CI, 0.73-0.91]), a finding consistent across external validation sets.

Conclusions and Relevance

In adult patients with septic shock receiving norepinephrine, the use of vasopressin was variable. A reinforcement learning model developed and validated in several observational datasets recommended more frequent and earlier use of vasopressin than average care patterns and was associated with reduced mortality.

Mortality time frame variability in septic shock clinical trials: A systematic review

OBJECTIVE

We sought to delineate the mortality outcome time frames reported in septic shock randomized control trials (RCTs).

DESIGN

Systematic review of PubMed, EMBASE, and the Cochrane Database of Systematic Reviews.

SETTING

Intensive care units.

PARTICIPANTS

Studies that included adult patients with septic shock.

INTERVENTIONS

Any type of intervention.

MAIN VARIABLES OF INTEREST

Information about the study, specific patient population, type of study intervention, specific intervention, and number of patients. Mortality time frames were analyzed for geographical differences and changes over time.

RESULTS

The search yielded 2660 unique citations. After screening, 132 eligible studies were identified. A total of 234 mortality time frames were collected from the included studies, of which 15 timeframes were unique. The most frequently reported time frame was 28-day mortality (n = 98, 74% of trials), followed by hospital mortality (n = 35, 27%), ICU mortality (n = 30, 23%), and 90-day mortality (n = 29, 22%). The most reported mortality time frame was 28 days in studies from every continent except Africa. The studies published between 2008 and 2013 (25%) more frequently reported hospital and ICU mortality combination than studies published between 2014 and 2019 (11.4%) (P = 0.043).

CONCLUSIONS

There was considerable variability in the mortality time frames reported in ICU-based septic shock trials. This variability may lead to under or overestimation of the problem, overlooking the effectiveness of the interventions studied, and further limiting the application of trials and their pooling in meta-analyses. A consensus regarding time frame reporting in septic shock trials is long overdue.

Identification of distinct clinical profiles of sepsis risk in paediatric emergency department patients using Bayesian profile regression

BACKGROUND

Sepsis affects 25 million children and neonates annually, causing significant mortality and morbidity. Early identification and treatment are crucial for improving outcomes. Identifying children at risk is challenging due to clinical heterogeneity and overlap with other conditions. Current evaluations of sepsis criteria adopt a variable-centred approach, evaluating each criterion independently. The objective of this study was to explore associations between patterns of sepsis screening criteria and sepsis risk in children screened in the emergency department (ED) to identify distinct profiles that describe the clinical heterogeneity of suspected sepsis.

METHODS

This secondary analysis involved 3473 children screened for sepsis across 12 EDs in Queensland, Australia. Bayesian profile regression was used to construct data-driven clinical profiles derived from sepsis screening criteria and their association with suspected sepsis, defined as senior medical officer diagnosis and antibiotic administration in the ED. Posterior risk probabilities (Prs) with 95% credible intervals (CIs) were calculated for each profile. Profiles were internally validated by assessing their association with sepsis, septic shock, organ dysfunction and infection sources, in both adjusted and unadjusted models.

RESULTS

Seven distinct clinical profiles were identified. Two profiles were labelled as high risk of suspected sepsis (profile 1, n=22: Pr 0.73, 95% CI 0.55, 0.89; profile 2, n=150: Pr 0.69, 95% CI 0.59, 0.80), four as moderate risk and one as low risk. High-risk profiles were characterised by severe illness indicators and elevated lactate levels. Moderate-risk profiles included criteria such as altered behaviour, young age (<3 months) and respiratory distress. High-risk profiles had strong associations with all clinical outcomes.

CONCLUSIONS

Seven clinical profiles were identified that varied in their risk of suspected sepsis and associated outcomes. Validation of these profiles in diverse populations and identification of which profiles are likely to benefit from certain interventions is needed.

Development of a machine learning-based predictive model for maxillary sinus cysts and exploration of clustering patterns

BACKGROUND AND OBJECTIVE

There are still many controversies about the factors influencing maxillary sinus cysts and their clinical management. This study aims to construct a prediction model of maxillary sinus cyst and explore its clustering pattern by cone beam computerized tomography (CBCT) technique and machine learning (ML) method to provide a theoretical basis for the prevention and clinical management of maxillary sinus cyst.

METHODS

In this study, 6000 CBCT images of maxillary sinus from 3093 patients were evaluated to document the possible influencing factors of maxillary sinus cysts, including gender, age, odontogenic factors, and anatomical factors. First, the characteristic variables were screened by multiple statistical methods, and ML methods were applied to construct a prediction model for maxillary sinus cysts. Second, the model was interpreted based on the SHapley Additive exPlanations (SHAP) values, and the risk of maxillary sinus cysts was predicted by generating a web page calculator. Finally, the K-mean clustering algorithm further identified risk factors for maxillary sinus cysts.

RESULTS

By comparing the various metrics in the training and test sets of multiple ML models, eXtreme Gradient Boosting (XGBoost) is the best model. The average area under curve (AUC) values of the XGBoost model in the training, validation, and test sets, respectively, are 0.939, 0.923, and 0.921, which indicates its excellent classification and discrimination ability. The cluster analysis model further categorized maxillary sinus cysts into high-risk and low-risk groups, with apical lesions, severe periodontitis, and age ≥ 53 as high-risk factors for maxillary sinus cysts.

CONCLUSION

These findings provide valuable insights into the etiology and risk stratification of maxillary sinus cysts, offering a theoretical basis for their prevention and clinical management. The integration of CBCT imaging and ML techniques holds the potential for prevention and personalized treatment strategies of maxillary sinus cysts.

Utilization of non-invasive ventilation before prehospital emergency anesthesia in trauma - a cohort analysis with machine learning

BACKGROUND

For preoxygenation, German guidelines consider non-invasive ventilation (NIV) as a possible method in prehospital trauma care in the absence of aspiration, severe head or face injuries, unconsciousness, or patient non-compliance. As data on the utilization and characteristics of patients receiving NIV are lacking, this study aims to identify predictors of NIV usage in trauma patients using machine learning and compare these findings with the current national guideline.

METHODS

A cross-regional registry of prehospital emergency services in southwestern Germany was searched for cases of emergency anesthesia in multiply injured patients in the period from 2018 to 2020. Initial vital signs, oxygen saturation, respiratory rate, heart rate, systolic blood pressure, Glasgow Coma Scale (GCS), injury pattern, shock index and age were examined using logistic regression. A decision tree algorithm was then applied in parallel to reduce the number of attributes, which were subsequently tested in several machine learning algorithms to predict the usage of NIV before the induction of anesthesia.

RESULTS

Of 992 patients with emergency anesthesia, 333 received NIV (34%). Attributes with a statistically significant influence (p < 0.05) in favour of NIV were bronchial spasm (odds ratio (OR) 119.75), dyspnea/cyanosis (OR 2.28), moderate and severe head injury (both OR 3.37) and the respiratory rate (OR 1.07). Main splitting points in the initial decision tree included auscultation (rhonchus and bronchial spasm), respiratory rate, heart rate, age, oxygen saturation and head injury with moderate head injury being more frequent in the NIV group (23% vs. 12%, p < 0.01). The rates of aspiration and the level of consciousness were equal in both groups (0.01% and median GCS 15, both p > 0.05). The prediction accuracy for NIV usage was high for all algorithms, except for multilayer perceptron and logistic regression. For instance, a Bayes Network yielded an AUC-ROC of 0.96 (95% CI, 0.95-0.96) and PRC-areas of 0.96 [0.96-0.96] for predicting and 0.95 [0.95-0.96] for excluding NIV usage.

CONCLUSIONS

Machine learning demonstrated an excellent categorizability of the cohort using only a few selected attributes. Injured patients without severe head injury who presented with dyspnea, cyanosis, or bronchial spasm were regularly preoxygenated with NIV, indicating a common prehospital practice. This usage appears to be in accordance with current German clinical guidelines. Further research should focus on other aspects of the decision making like airway anatomy and investigate the impact of preoxygenation with NIV in prehospital trauma care on relevant outcome parameters, as the current evidence level is limited.

Distinct medical and substance use histories associate with cognitive decline in Alzheimer's disease

INTRODUCTION

Phenotype clustering reduces patient heterogeneity and could be useful when designing precision clinical trials. We hypothesized that the onset of early cognitive decline in patients would exhibit variance predicated on the clinical history documented prior to an Alzheimer's disease (AD) diagnosis.

METHODS

Self-reported medical and substance use history (i.e., problem history) was used to cluster participants from the National Alzheimer's Coordinating Center (NACC) into distinct subtypes. Linear mixed effects modeling was used to determine the effect of problem history subtype on cognitive decline over 2 years.

RESULTS

Two thousand seven hundred fifty-four individuals were partitioned into three subtypes: minimal (n = 1380), substance use (n = 1038), and cardiovascular (n = 336). The cardiovascular problem history subtype had significantly worse cognitive decline over a 2 year follow-up period (p = 0.013).

DISCUSSION

Our study highlights the need to account for problem history to reduce heterogeneity of outcomes in AD clinical trials.

HIGHLIGHTS

Clinical data were used to identify subtypes of patients with Alzheimer's disease (AD) in the National Alzheimer's Coordinating Center dataset. Three problem history subtypes were found: minimal, substance use, and cardiovascular. The mean change in Clinical Dementia Rating Sum of Boxes (CDR-SB) was assessed over a 2 year follow-up. The cardiovascular subtype was associated with the worst cognitive decline. The magnitude of change in CDR-SB was similar to recent AD clinical trials.

Sepsis in silico: definition, development and application of an electronic phenotype for sepsis

Repurposing electronic health record (EHR) or electronic medical record (EMR) data holds significant promise for evidence-based epidemic intelligence and research. Key challenges include sepsis recognition by physicians and issues with EHR and EMR data. Recent advances in data-driven techniques, alongside initiatives like the Surviving Sepsis Campaign and the Severe Sepsis and Septic Shock Management Bundle (SEP-1), have improved sepsis definition, early detection, subtype characterization, prognostication and personalized treatment. This includes identifying potential biomarkers or digital signatures to enhance diagnosis, guide therapy and optimize clinical management. Machine learning applications play a crucial role in identifying biomarkers and digital signatures associated with sepsis and its sub-phenotypes. Additionally, electronic phenotyping, leveraging EHR and EMR data, has emerged as a valuable tool for evidence-based sepsis identification and management. This review examines methods for identifying sepsis cohorts, focusing on two main approaches: utilizing health administrative data with standardized diagnostic coding via the International Classification of Diseases and integrating clinical data. This overview provides a comprehensive analysis of current cohort identification and electronic phenotyping strategies for sepsis, highlighting their potential applications and challenges. The accuracy of an electronic phenotype or signature is pivotal for precision medicine, enabling a shift from subjective clinical descriptions to data-driven insights.

Bundle compliance patterns in septic shock and their association with patient outcomes: an unsupervised cluster analysis

Adhering to bundle-based care recommendations within stringent time constraints presents a profound challenge. Elements within these bundles hold varying degrees of significance. We aimed to evaluate the Surviving Sepsis Campaign (SSC) hour-one bundle compliance patterns and their association with patient outcomes. Utilizing the Medical Information Mart for Intensive Care-IV 1.0 dataset, this retrospective cohort study evaluated patients with sepsis who developed shock and were admitted to the intensive care unit between 2008 and 2019. The execution of five hour-one bundle interventions were assessed. Patients with similar treatment profiles were categorized into clusters using unsupervised machine learning. Primary outcomes included in-hospital and 1-year mortality. Four clusters were identified: C#0 (n = 4716) had the poorest bundle compliance. C#1 (n = 1117) had perfect antibiotic adherence with modest fluid and serum lactate measurement adherence. C#2 (n = 850) exhibited full adherence to lactate measurement and low adherence to fluid administration, blood culture, and vasopressors, while C#3 (n = 381) achieved complete adherence to fluid administration and the highest adherence to vasopressor requirements in the entire cohort. Adjusting for covariates, C#1 and C#3 were associated with reduced odds of in-hospital mortality compared to C#0 (adjusted odds ratio [aOR] = 0·83; 95% confidence interval [CI] 0·7-0·97 and aOR = 0·7; 95% CI 0·53-0·91, respectively). C#1 exhibited significantly better 1-year survival (adjusted hazard ratio [aHR] = 0·9; 95%CI 0·81-0·99). We were able to identify distinct clusters of SSC hour-one bundle adherence patterns using unsupervised machine learning techniques, which were associated with patient outcomes.

Profiling of Cardiogenic Shock: Incorporating Machine Learning Into Bedside Management

Cardiogenic shock (CS) is a complex clinical syndrome with various etiologies and clinical presentations. Despite advances in therapeutic options, mortality remains high, and clinical trials in the field are complicated in part by the heterogeneity of CS patients. More individualized targeted therapeutic approaches might improve outcomes in CS, but their implementation remains challenging. The present review discusses current and emerging machine learning-based approaches, including unsupervised and supervised learning methods that use real-world clinical data to individualize therapeutic strategies for CS patients. We will discuss the rationale for each approach, potential advantages and disadvantages, and how these strategies can inform clinical trial design and management decisions.

Unsupervised Classification of the Host Response Identifies Dominant Pathobiological Signatures of Sepsis in Sub-Saharan Africa

Rationale: The global burden of sepsis is concentrated in sub-Saharan Africa, where inciting pathogens are diverse and HIV coinfection is a major driver of poor outcomes. Biological heterogeneity inherent to sepsis in this setting is poorly defined. Objectives: To identify dominant pathobiological signatures of sepsis in sub-Saharan Africa and their relationship to clinical phenotypes, patient outcomes, and biological classifications of sepsis identified in high-income countries (HICs). Methods: We analyzed two prospective cohorts of adults hospitalized with sepsis (severe infection with quick Sepsis-related Organ Failure Assessment score ⩾1) at disparate settings in Uganda (discovery cohort [Entebbe, urban], n = 242; validation cohort [Tororo, rural], n = 253). To identify pathobiological signatures in the discovery cohort, we applied unsupervised clustering to 173 soluble proteins reflecting key domains of the host response to severe infection. A random forest-derived classifier was used to predict signature assignment in the validation cohort. Measurements and Main Results: Two signatures (Uganda Sepsis Signature [USS]-1 and USS-2) were identified in the discovery cohort, distinguished by expression of proteins involved in myeloid cell and inflammasome activation, T-cell costimulation and exhaustion, and endothelial barrier dysfunction. A five-protein classifier (area under the receiver operating characteristic curve, 0.97) reproduced two signatures in the validation cohort with similar biological profiles. In both cohorts, USS-2 mapped to a more severe clinical phenotype associated with HIV and related immunosuppression, severe tuberculosis, and increased risk of 30-day mortality. Substantial biological overlap was observed between USS-2 and hyperinflammatory and reactive sepsis phenotypes identified in HICs. Conclusions: We identified prognostically enriched pathobiological signatures among patients with sepsis with diverse infections and high HIV prevalence in Uganda. Globally inclusive investigations are needed to define generalizable and context-specific mechanisms of sepsis pathobiology, with the goal of improving access to precision medicine treatment strategies.