Impact of an Antibiotic-specific Sepsis Bundle on Appropriate and Timely Antibiotic Administration for Severe Sepsis in the Emergency Department

Interpretable machine learning for predicting sepsis risk in emergency triage patients

The study aimed to develop and validate a sepsis prediction model using structured electronic medical records (sEMR) and machine learning (ML) methods in emergency triage. The goal was to enhance early sepsis screening by integrating comprehensive triage information beyond vital signs. This retrospective cohort study utilized data from the MIMIC-IV database. Two models were developed: Model 1 based on vital signs alone, and Model 2 incorporating vital signs, demographic characteristics, medical history, and chief complaints. Eight ML algorithms were employed, and model performance was evaluated using metrics such as AUC, F1 Score, and calibration curves. SHapley Additive exPlanations (SHAP) and Local Interpretable Model-agnostic Explanations (LIME) methods were used to enhance model interpretability. The study included 189,617 patients, with 5.95% diagnosed with sepsis. Model 2 consistently outperformed Model 1 across most algorithms. In Model 2, Gradient Boosting achieved the highest AUC of 0.83, followed by Extra Tree, Random Forest, and Support Vector Machine (all 0.82). The SHAP method provided more comprehensible explanations for the Gradient Boosting algorithm. Modeling with comprehensive triage information using sEMR and ML methods was more effective in predicting sepsis at triage compared to using vital signs alone. Interpretable ML enhanced model transparency and provided sepsis prediction probabilities, offering a feasible approach for early sepsis screening and aiding healthcare professionals in making informed decisions during the triage process.

Early predictors of bacterial pneumonia infection in children with congenital heart disease after cardiopulmonary bypass: a single-centre retrospective study

OBJECTIVES

The objective of this study was to evaluate the early predictors of bacterial pneumonia infection in children with congenital heart disease (CHD) after cardiopulmonary bypass (CPB).

DESIGN

Retrospective study.

SETTING

A freestanding tertiary paediatric hospital in China.

PARTICIPANTS

Patients admitted to the hospital due to CHD who underwent open-heart surgery.

OUTCOME MEASURES

We retrospectively reviewed and analysed data from 1622 patients with CHD after CPB from June 2018 to December 2020 at the Children's Hospital of Nanjing Medical University. Enrolled patients were assigned to an infection group or a non-infection group according to the presence of postoperative bacterial pneumonia infection, and the differences in clinical indicators were compared. Potential predictors were analysed by multivariate logistic regression analysis and area under the curve (AUC) analysis.

RESULTS

Among the 376 patients (23.2%) in the infection group, the three most common bacteria were Streptococcus pneumoniae in 67 patients (17.8%), Escherichia coli in 63 patients (16.8%) and Haemophilus influenzae in 53 patients (14.1%). The infection group exhibited a lower weight (8.0 (6.0-11.5) kg vs 11.0 (7.5-14.5) kg, p<0.001). In the infection group, procalcitonin (PCT) (ng/mL: 4.72 (1.38-9.52) vs 1.28 (0.47-3.74), p<0.001) and C reactive protein (CRP) (mg/L: 21.0 (12.1-32.0) vs 17.0 (10.0-27.0), p<0.001) levels were significantly greater than those in the non-infection group. Binary logistic regression analysis revealed that weight, PCT and CRP were independent risk factors for pulmonary bacterial infection after CPB. The AUCs of weight, PCT, CRP and PCT+CRP for predicting pulmonary bacterial infection after CPB were 0.632 (95% CI 0.600 to 0.664), 0.697 (95% CI 0.667 to 0.727), 0.586 (95% CI 0.554 to 0.618) and 0.694 (95% CI 0.664 to 0.724), respectively, and the cut-off values were ≤10.25 kg, ≥4.25 ng/mL, ≥6.50 mg/L and ≥0.20, respectively. The sensitivities were 69.7%, 54.0%, 93.9% and 70.2%, and the specificities were 53.5%, 77.7%, 19.4% and 59.1%, respectively.

CONCLUSIONS

In our study, weight, PCT and CRP were found to be independent predictors of pulmonary bacterial infection after CPB. Moreover, PCT was the most specific predictor, and CRP was the most sensitive independent predictor that might be beneficial for the early diagnosis of pulmonary bacterial infection after CPB in patients with CHD.

Impact of emergency medicine clinical pharmacist practitioner-driven sepsis antibiotic interventions

BACKGROUND

The 2021 Surviving Sepsis Campaign Guidelines recommend administration of antimicrobials within the first hour of recognition of sepsis. Over the last decade, several studies have demonstrated improved time-to-antibiotic administration and antibiotic appropriateness when a pharmacist was involved in the care of patients with sepsis. To our knowledge, no studies evaluating the appropriate use of antibiotics in sepsis driven entirely by an Emergency Medicine (EM) Clinical Pharmacist Practitioner (CPP) have been published. The purpose of this study is to evaluate the impact of an EM CPP-driven protocol on antimicrobial interventions in patients with sepsis in the emergency department (ED).

METHODS

This was a retrospective comparison of patients with sepsis for whom antimicrobials were ordered in the ED without pharmacist intervention to patients whose antimicrobials were ordered by an EM CPP via a sepsis consult to pharmacy. An EM CPP reviewed individual patient profiles for pertinent historical admissions, culture data, and allergy profiles to guide antimicrobial selection for the suspected source of infection and entered orders under their scope of practice with formal documentation in the electronic medical record (EMR). The primary objective of this study was to compare the rates of appropriate empiric antibiotic utilization in septic patients admitted from the ED pre- and post-protocol implementation. Secondary endpoints included the following, broadening of ED-initiated empiric antibiotics on hospital admission, time-to-antibiotic administration, in-hospital mortality, Rapid Emergency Medicine Score (REMS) association with in-hospital mortality, and hospital length of stay.

RESULTS

A total of 144 patients were included: 80 patients prescribed antibiotics without pharmacist intervention and 64 prescribed antibiotics by an EM CPP. Appropriate empiric antibiotic selection in the ED improved from 57.5% (46/80) to 86% (55/64) with EM CPP intervention (difference 28.5%; p < 0.01). Time-to-first antibiotic administration decreased by 64 min (p < 0.01). Administration of antibiotics within 60 min, broadening of antibiotics on admission, hospital length of stay, and in-hospital mortality did not significantly differ across groups.

CONCLUSIONS

In this small, single-center study, an EM Clinical Pharmacist Practitioner-driven protocol for patients with sepsis in the emergency department improved the rate of appropriate empiric antimicrobial selection and time-to-antibiotic administration.

Association between the timing of antibiotics administration and outcome of neonatal sepsis

BACKGROUND

Sepsis is a life-threatening condition that requires aggressive and rapid intervention. However, data on the association between antibiotic administration timing in neonatal sepsis and neonatal outcomes is limited, particularly in the gulf area.

OBJECTIVE

This study aimed to evaluate the association between the timing of antibiotic administration and the outcome of neonatal sepsis.

DESIGN AND SETTING

This retrospective comparative study was conducted through data collection from medical records of patients with neonatal sepsis. The patients were categorized into two groups based on the time interval between antibiotic prescription and drug administration: non-delayed group consisted of patients who received antibiotics within 3 h and the delayed group consisted of those who received antibiotics after 3 h.

RESULTS

A total of 237 neonates diagnosed with sepsis were included, of which 9.3% had necrotizing enterocolitis, 35% had bronchopulmonary dysplasia, and 6.3% had maternal chorioamnionitis. Additionally, 18.6% of the neonates' mothers were prescribed with antibiotics during labor, and 5.5% had maternal fever. Staphylococcus epidermidis was the most commonly isolated strain (24.1%). Of the total neonates, 87.3% received antibiotics within the first 3 h from the prescription. Survival rate was significantly higher and the risk of complications such as necrotizing enterocolitis, and bronchopulmonary dysplasia was significantly lower in the non-delayed group. Moreover, the length of hospital stay was significantly shorter in the non-delayed group.

CONCLUSION

Early antibiotic administration in patients with neonatal sepsis can improve the survival rate and reduce the incidence of complications.

Clinical and economical improvements after introducing rapid identification of bacteria and early antibiotic susceptibility testing in sepsis and bloodstream infections. Results of the PHENOMENON study

Background: Sepsis and bloodstream infections pose severe challenges in intensive care. Early reliable diagnosis is the key to successful therapy. The objective of the study presented here was to investigate the clinical and economical effects of the new Pheno^TM BC test, which allows bacteria identification (ID) and antimicrobial susceptibility testing (AST) in approximately 7 hours after a blood culture becomes positive (BC+). Methods: Historically controlled interventional study. Population: patients with BC+ and ICU admission. Inadequate initial antimicrobial therapy (IAT) is need of therapy change based on result. Prospectively the new test was used in addition. Primary endpoint: time-to-result in hours. Contribution margin (CM) i.e. revenue - costs was computed. All patients formed the intention-to-treat population (ITT). Patients with complete cost data formed the modified ITT group (mITT). CM results were calculated for mITT and PP. Further analyses: length-of-stay (LOS) and mortality. Results: 223 historical and 200 prospective patients were included. Time to result (ITT) was shortened by 51.1 hours (83 vs. 31.9; p<0.001). Overall savings (mITT) were 257,100 € (-301,264 € vs. -44,164 €). 143 of 181 (79%) patients had a test performed, 126 of 143 (88%) having a clinically useable result. 40 (32%) had IAT vs. 65 (29%) in the historic cohort. Median time to AST in PP was shortened by 61.7 hours (89.5 vs. 27.8; p<0.001). LOS was shortened 7 days (28 vs. 19; p=0.226) and mortality was 8% (40.5% vs. 32.5%; p=0.440) lower. Median CM +3,074.80 € per case (-2,350.50 € vs. +724.70 €; p=0.040). Conclusion: The new Pheno^TM ID+AST test leads to faster and clinically meaningful results and saves money by shortening LOS on the ICU.

The influence of nurse allocated triage category on the care of patients with sepsis in the emergency department: A retrospective review

BACKGROUND

'Gold standard' sepsis care encompasses the recognition and treatment of sepsis within one hour of emergency department (ED) presentation. Early treatment of patients with sepsis reduces mortality. The aim of this study was to examine the effect that the nurse allocated Australasian Triage Scale (ATS) Category allocation had on ED patient treatment for severe sepsis and septic shock.

METHODS

A two-year retrospective observational cohort study from a single major metropolitan ED, including all patients with severe sepsis or septic shock.

RESULTS

Sixty patients were included in this study. Sepsis was recognised at triage for the majority of patients (n=38, 63%), and most were allocated an ATS Category Two (n=39). Almost half of the patients received all elements of the sepsis bundle within one hour of arrival (n=27,45%). Patients allocated an ATS Category One or Two had a shorter time to lactate collection (p=0.003), blood culture procurement (p=0.009) and intravenous antibiotic administration (p=0.021) compared with patients who were allocated ATS Category Three or Four.

CONCLUSIONS

Most patients presenting with sepsis were recognised by the triage nurse and allocated a high acuity ATS category accordingly. As sepsis is a time-critical condition and a high acuity triage allocation reduces time to treatment, we recommend all Australian EDs should implement a standard approach to sepsis triage by allocating an ATS Category of One or Two to all patients suspected of having sepsis, thus reflecting the urgency of their disease.

[Anti-infective treatment : Treatment strategies for sepsis and septic shock]

Sepsis and septic shock are still associated with a high mortality and morbidity. A decisive factor for improvement of the outcome is the prompt initiation of an effective antibiotic treatment. The early recognition of sepsis within the first hour is here one of the biggest challenges. Effective empirical treatment comprises purposefully selected broad-spectrum antibiotics and also combination treatment or antimycotics in special situations. De-escalation strategies to narrow down or shorten the treatment are safe and can limit the side effects.

BundlED Up: A Narrative Review of Antimicrobial Stewardship Initiatives and Bundles in the Emergency Department

Antimicrobial stewardship (ASP) is becoming an increasingly high priority worldwide, yet the emergency department (ED) is an area where stewardship is often neglected. Implementing care bundles, guidelines, and protocols appears to be a rational strategy for ED stewardship given the inherently dynamic and hectic environment of care. Multiple questions still exist such as whether to target certain disease states, optimal implementation of ASP interventions in the ED, and the benefit of unique ED-specific guidelines and protocols. A narrative review was performed on interventions, guidelines, and bundles implemented in the ED setting, in an effort to improve ASP or management of infectious diseases. This review is meant to serve as a framework for the reader to implement these practices at their own institution. We examined various studies related to ASP interventions or care bundles in the ED which included: CNS infections (one study), skin and soft-tissue infections (one study), respiratory infections (four studies), urinary tract infections and sexually transmitted infections (eight studies), sepsis (two studies), culture follow-up programs (four studies), and stewardship in general or multiple infection types (five studies). The interventions in this review were diverse, yet the majority showed a benefit in clinical outcomes or a decrease in antimicrobial use. Care bundles, guidelines, and antimicrobial stewardship interventions can streamline care and improve the management of common infectious diseases seen in the ED.

Time Spent in the Emergency Department and Outcomes in Patients With Severe Sepsis and Septic Shock

A majority of patients with severe sepsis and septic shock are first evaluated in the emergency department (ED). Methods such as screening tools have proven advantageous in earlier identification, allowing for timely initiation of treatment. Delay in symptom presentation and ED overcrowding contribute to deferment of sepsis bundle components and admission. To examine the impact of time from ED arrival to inpatient admission on mortality and length of stay (LOS) in patients with severe sepsis or septic shock. A retrospective analysis of adult patients with severe sepsis or septic shock was completed for those presenting between January 2013 and December 2014. Patients were dichotomized on the basis of the length of time from completed triage in the ED to intensive care unit (ICU) admission (at less than 6 hr and at 6 hr or more). Of the 294 patients screened, 172 patients (58.5%) met inclusion criteria (n = 70 cases at less than 6 hr; n = 102 at 6 hr or more). Mean wait time from ED arrival to ICU admission was 470.7 ± 333.9 min (range = 84-2,390 min). Groups were similar in baseline, disease severity, and bundle characteristics. There were no differences in the less than 6-hr group compared with the 6-hr-or-more group in rates of 30-day mortality (37.1% vs. 32.4%; p = 0.52), as well as in-hospital (27.1% vs. 23.5%; p = 0.59) or 90-day mortality (42.9% vs. 34.3%; p = 0.26). There were also no differences in hospital or ICU LOS. Timing of transfer from the ED to the ICU was not found to impact mortality or LOS. These results suggest that the ED can provide similar sepsis care to that in the ICU when transfer is delayed in patients with sepsis.

Is Procalcitonin Useful in Pediatric Critical Care Patients?

This review examines the use of procalcitonin in different clinical situations in the pediatric patient, with special emphasis on those requiring intensive care. We review the latest articles on its potency as a biomarker in both infectious processes at diagnosis and on the response to treatment.

Appropriate Antibiotic Therapy

Prescribing antibiotics is an essential component of initial therapy in sepsis. Early antibiotics are an important component of therapy, but speed of administration should not overshadow the patient-specific characteristics that determine the optimal breadth of antimicrobial therapy. Cultures should be drawn before antibiotic therapy if it does not significantly delay administration. Combination antibiotic therapy against gram-negative infections is not routinely required, and combination therapy involving vancomycin and piperacillin/tazobactam is associated with an increase in acute kidney injury. Emergency practitioners should be aware of special considerations in the administration and dosing of antibiotics in order to deliver optimal care to septic patients.

Metagenomic analysis of bloodstream infections in patients with acute leukemia and therapy-induced neutropenia

Leukemic patients are often immunocompromised due to underlying conditions, comorbidities and the effects of chemotherapy, and thus at risk for developing systemic infections. Bloodstream infection (BSI) is a severe complication in neutropenic patients, and is associated with increased mortality. BSI is routinely diagnosed with blood culture, which only detects culturable pathogens. We analyzed 27 blood samples from 9 patients with acute leukemia and suspected BSI at different time points of their antimicrobial treatment using shotgun metagenomics sequencing in order to detect unculturable and non-bacterial pathogens. Our findings confirm the presence of bacterial, fungal and viral pathogens alongside antimicrobial resistance genes. Decreased white blood cell (WBC) counts were associated with the presence of microbial DNA, and was inversely proportional to the number of sequencing reads. This study could indicate the use of high-throughput sequencing for personalized antimicrobial treatments in BSIs.