Background: Neonatal sepsis is one of the major causes of morbidity and mortality in newborns. However, atypical clinical manifestations and symptoms make the early diagnosis of neonatal sepsis a challenge. Relatively high-serum soluble urokinase-type plasminogen activator receptor (suPAR) has been implicated as a diagnostic biomarker for adult sepsis. Therefore, the meta-analysis is intended to explore the diagnostic value of suPAR for neonatal sepsis.

Methods: The PubMed, Cochrane Library, Embase, Web of Science, China National Knowledge Infrastructure, China Biological Medicine Disk, and Wanfang databases were retrieved from inception to 31 December 2022 to collect diagnostic accuracy studies about suPAR for neonatal sepsis. Two reviewers independently screened the literature, extracted data, and assessed the risk of bias in the included studies using the quality assessment of diagnostic accuracy studies-2 (QUADAS-2) tool. Then, a meta-analysis was performed using Stata 15.0 software.

Results: A total of six articles involving eight studies were included. The results of the meta-analysis showed that the pooled sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, and diagnostic odds ratio were 0.89 [95%CI (0.83–0.93)], 0.94 [95%CI (0.77–0.98)], 14 [95%CI (3.5–55.2)], 0.12 [95%CI (0.08–0.18)], and 117 [95%CI (24–567)], respectively. The area under the curve (AUC) of summary receiver operator characteristic (SROC) curves was 0.92 [95%CI (0.90–0.94)]. Sensitivity analysis confirmed the stability of the results, and publication bias was not observed. Fagan’s nomogram results demonstrated the clinical availability of the findings.

Conclusion: Current evidence suggests that suPAR has potential diagnostic value for neonatal sepsis. Owing to the limited quality of the included studies, more high-quality studies are needed to verify the above conclusion.

Background: Patients admitted to hospital with sepsis are at persistent risk of poor outcome after discharge. Many tools are available to risk-stratify sepsis patients for in-hospital mortality. This study aimed to identify the best risk-stratification tool to prognosticate outcome 180 days after admission via the emergency department (ED) with suspected sepsis.

Methods: A retrospective observational cohort study was performed of adult ED patients who were admitted after receiving intravenous antibiotics for the treatment of a suspected sepsis, between 1^st March and 31^st August 2019. The Risk-stratification of ED suspected Sepsis (REDS) score, SOFA score, Red-flag sepsis criteria met, NICE high-risk criteria met, the NEWS2 score and the SIRS criteria, were calculated for each patient. Death and survival at 180 days were noted. Patients were stratified in to high and low-risk groups as per accepted criteria for each risk-stratification tool. Kaplan–Meier curves were plotted for each tool and the log-rank test performed. The tools were compared using Cox-proportional hazard regression (CPHR). The tools were studied further in those without the following specified co-morbidities: Dementia, malignancy, Rockwood Frailty score of 6 or more, long-term oxygen therapy and previous do-not-resuscitate orders.

Results: Of the 1,057 patients studied 146 (13.8%) died at hospital discharge and 284 were known to have died within 180 days. Overall survival proportion was 74.4% at 180 days and 8.6% of the population was censored before 180 days. Only the REDS and SOFA scores identified less than 50% of the population as high-risk. All tools except the SIRS criteria, prognosticated for outcome at 180 days; Log-rank tests between high and low-risk groups were: REDS score p < 0.0001, SOFA score p < 0.0001, Red-flag criteria p = 0.001, NICE high-risk criteria p = 0.0001, NEWS2 score p = 0.003 and SIRS criteria p = 0.98. On CPHR, the REDS [Hazard ratio (HR) 2.54 (1.92–3.35)] and SOFA [HR 1.58 (1.24–2.03)] scores out-performed the other risk-stratification tools. In patients without the specified co-morbidities, only the REDS score and the SOFA score risk-stratified for outcome at 180 days.

Conclusion: In this study, all the risk-stratification tools studied were found to prognosticate for outcome at 180 days, except the SIRS criteria. The REDS and SOFA scores outperformed the other tools.

Purpose: To build machine learning models for predicting the risk of in-hospital death in patients with sepsis within 48 h, using only dynamic changes in the patient's vital signs.

Methods: This retrospective observational cohort study enrolled septic patients from five emergency departments (ED) in Taiwan. We adopted seven variables, i.e., age, sex, systolic blood pressure, diastolic blood pressure, heart rate, respiratory rate, and body temperature.

Results: Among all 353,253 visits, after excluding 159,607 visits (45%), the study group consisted of 193,646 ED visits. With a leading time of 6 h, the convolutional neural networks (CNNs), long short-term memory (LSTM), and random forest (RF) had accuracy rates of 0.905, 0.817, and 0.835, respectively, and the area under the receiver operating characteristic curve (AUC) was 0.840, 0.761, and 0.770, respectively. With a leading time of 48 h, the CNN, LSTM, and RF achieved accuracy rates of 0.828, 0759, and 0.805, respectively, and an AUC of 0.811, 0.734, and 0.776, respectively.

Conclusion: By analyzing dynamic vital sign data, machine learning models can predict mortality in septic patients within 6 to 48 h of admission. The performance of the testing models is more accurate if the lead time is closer to the event.