Background: Sepsis is an inflammatory infectious state with significant mortality and healthcare system burden. Early detection and management dramatically improves patient outcomes. As the entry point for many, emergency departments (EDs) are ideally placed to initiate sepsis treatment and determine admission need. Cell-free DNA (cfDNA) is a potential predictor of outcomes in ED sepsis patients; previous intensive care unit (ICU) studies have correlated high cfDNA with in-hospital mortality. During a 6 month study period at a large academic ED, 453 consecutive blood samples were collected prospectively from septic patients (study samples drawn as part of standard sepsis bundle), with additional retrospective chart review; of these, 140 patients met inclusion criteria (samples processed within 24 hours; no immunocompromise, no active cancer, no hospice status), had cfDNA extracted, and were risk-stratified to high (n=90) and low risk (n=50) based on disposition (high risk: admission ≥24 hrs). Variables were modeled by logistic regression. Scores from other illness prediction algorithms were modeled including qSOFA (Quick Sequential Organ Failure Assessment), REMS (Rapid Emergency Medicine Score), MEDS (Mortality in Emergency Department Sepsis), and CURB-65 (Confusion, Uremia, Respiratory rate, Blood pressure, age ≥65).
No single variable was a sound predictor of disposition, though a multivariate model including cfDNA (AUC 0.82) was better than other standardized models at predicting need for admission. Multivariate model: age ≥55 years; cfDNA ≥ 80 ng/mL; lactate ≥2.2 mg/dL; maximum heart rate ≥120 bpm. In comparison qSOFA had relatively strong predictive ability (AUC 0.76), while remaining models and algorithms fared poorly.
Conclusion: A multivariate model may translate to a risk stratification score specific to disposition planning, rather than predicting in-hospital mortality as has been found in ICU studies. However, a decent scoring algorithm (qSOFA) already exists which does not require the time or expense of laboratory analysis. cfDNA is potentially an exciting marker of illness severity, but further evaluation is needed to determine if the additive clinical benefit of cfDNA above qSOFA justifies its use.