Daily Sepsis Research Analysis
Three studies advance sepsis care across stewardship, diagnostics, and precision glycemic risk stratification. A prospective NICU stewardship program reduced 14- and 28-day mortality, an interpretable MIMIC-IV cohort showed metabolic state–specific mortality risk from combined stress hyperglycemia ratio and glucose variability, and implementation of BCID2 multiplex PCR in bacteremia was associated with lower in-hospital mortality.
Summary
Three studies advance sepsis care across stewardship, diagnostics, and precision glycemic risk stratification. A prospective NICU stewardship program reduced 14- and 28-day mortality, an interpretable MIMIC-IV cohort showed metabolic state–specific mortality risk from combined stress hyperglycemia ratio and glucose variability, and implementation of BCID2 multiplex PCR in bacteremia was associated with lower in-hospital mortality.
Research Themes
- Antibiotic stewardship and outcomes in neonatal sepsis
- Precision glycemic phenotyping for sepsis prognosis
- Rapid molecular diagnostics for bacteremia and sepsis care
Selected Articles
1. Clinical impact of an antibiotic stewardship program in a neonatal intensive care unit at a tertiary care hospital: a prospective quasi-experimental clinical study.
In a prospective quasi-experimental study of 1200 NICU patients, implementing a locally tailored antibiotic stewardship program—including NICU-specific antibiograms and a revised neonatal sepsis protocol—improved prescribing and significantly reduced 14- and 28-day mortality. The program increased appropriate diagnostics (e.g., CRP testing) alongside protocolized therapy.
Impact: Prospective, real-world stewardship implementation with mortality benefit in neonatal sepsis is rare and highly actionable. It demonstrates that local antibiograms and protocolized care can translate to survival gains.
Clinical Implications: Hospitals should implement NICU-specific antibiotic stewardship with local antibiograms and protocolized neonatal sepsis regimens, including diagnostic stewardship (e.g., CRP), to reduce short-term mortality while combating resistance.
Key Findings
- Prospective quasi-experimental NICU stewardship (n=1200) reduced 14- and 28-day mortality.
- Program included NICU-specific antibiograms and protocolized neonatal sepsis therapy aligned with local susceptibility patterns.
- Diagnostic stewardship increased (e.g., CRP testing frequency rose), supporting targeted antibiotic use.
Methodological Strengths
- Prospective quasi-experimental design with predefined stewardship measures and trial registration (NCT04039152).
- Large NICU cohort with phase-wise evaluation and protocolized implementation.
Limitations
- Single-center, nonrandomized design may be subject to secular trends and confounding.
- Abstract does not report effect sizes for all outcomes; full data needed for external benchmarking.
Future Directions: Cluster-randomized or stepped-wedge trials across NICUs to confirm mortality benefits, assess antimicrobial resistance trajectories, and evaluate unintended consequences (e.g., fungal infections).
2. Simultaneous Assessment of Stress Hyperglycemia Ratio and Glucose Variability to Predict All-cause Mortality in Sepsis Patients Across Different Glucose Metabolic States: an Observational Cohort Study With Interpretable Machine Learning Approach.
In 4,838 sepsis patients from MIMIC-IV, combined SHR and GV stratified 28-day mortality differentially by glucose metabolic state: high SHR/high GV in NGR, low SHR/high GV in Pre-DM, and high SHR/low GV in DM carried the greatest risks. Interpretable ML models achieved AUCs up to 0.776, highlighting a precision glycemic risk framework.
Impact: This study operationalizes precision glycemic phenotyping in sepsis, showing that the prognostic impact of SHR and GV depends on baseline glucose metabolism. It provides actionable risk markers and interpretable models to guide tailored glycemic management.
Clinical Implications: Monitor both SHR and GV and interpret them in the context of baseline glucose status (NGR, Pre-DM, DM) to risk-stratify sepsis patients. These markers could inform targets and aggressiveness of glycemic control pending prospective validation.
Key Findings
- In 4,838 sepsis patients, the combination of high SHR (>1.23) and high GV (>28.56) conferred the highest 28-day mortality risk in NGR (HR 2.06; 95% CI 1.40–3.04).
- Pre-DM patients with low SHR/high GV had the greatest 28-day mortality risk (HR 2.45; 95% CI 1.73–3.48); DM patients with high SHR/low GV had the highest risk (HR 1.46; 95% CI 1.06–2.01).
- Interpretable ML models (RF, LR AUC 0.776; XGBoost AUC 0.746) identified SHR and GV as key predictors across metabolic strata.
Methodological Strengths
- Large, well-characterized ICU cohort (MIMIC-IV) with robust survival analyses (KM, Cox, RCS, landmark).
- Model transparency via SHAP enhances interpretability and clinical translation.
Limitations
- Retrospective single-database analysis with potential residual confounding and measurement biases.
- No external prospective validation; interventional implications for glycemic control remain to be tested.
Future Directions: Prospective, multi-center validation and randomized trials to test SHR/GV-guided glycemic strategies by metabolic phenotype.
3. Impact of multiplex polymerase chain reaction testing in patients with bacteremia.
In a retrospective pre-post study of 2,872 bacteremic inpatients, implementing the FilmArray BCID2 panel was associated with significantly lower in-hospital mortality (aOR 0.58, 95% CI 0.37–0.92). Outcomes included LOS, DOT, and DASC, supporting the clinical utility of rapid multiplex PCR to optimize therapy.
Impact: Real-world evidence links rapid multiplex molecular diagnostics to survival gains in bacteremia, a cornerstone of sepsis care, supporting broader BCID2 adoption integrated with stewardship.
Clinical Implications: Integrate BCID2 with antimicrobial stewardship to shorten time-to-targeted therapy, reduce broad-spectrum exposure (DASC), and potentially improve survival in bacteremia and sepsis pathways.
Key Findings
- Among 2,872 positive blood culture patients, BCID2 implementation was associated with reduced in-hospital mortality (aOR 0.58; 95% CI 0.37–0.92).
- Pre- vs post-implementation comparison assessed LOS, DOT, and DASC alongside mortality.
- Findings support rapid pathogen and resistance gene detection to optimize antimicrobial therapy.
Methodological Strengths
- Large pre-post cohort with adjusted analyses linking implementation to outcomes.
- Evaluated multiple clinically relevant endpoints (mortality, LOS, DOT, DASC).
Limitations
- Single-center retrospective design; potential secular trends and unmeasured confounders.
- Abstract truncation limits visibility into effect sizes for LOS, DOT, and DASC.
Future Directions: Prospective multicenter implementation studies with time-to-therapy metrics, cost-effectiveness, and integration with stewardship protocols.