Precocious sepsis detection in critical care patients using heart rate approximate entropy

Abstract

Sepsis is diagnosed when symptoms and signs are installed. Diagnostic delays cause high morbidity and mortality. Heart rate variability (HRV) has been proposed as a sepsis biomarker. Approximate Entropy (ApEn) is a measure of the complexity of the HRV signal. Objective: To determine heart rate ApEn value for precocious sepsis detection.

A prospective observational study was carried out at Hospital Misericordia, between 2016 - 2018. Both gender patients >15 years old admitted to ICU were included. Patients with disorder that could alter HRV (cardiopulmonary, neurological, arrhythmias, beta-blockers, or calcium blockers) were excluded. Study variable: blood culture confirmed sepsis. Main predictor: hearth rate (HR) ApEn. HR was determined by pulse and continuous ECG monitoring. Covariates: age, sex, APACHEII, Glasgow and SOFA scores, morbidity, admission diagnosis, complications, and mortality. Discrete variables were described in absolute and relative frequencies with 95%CI. Normal continuous variables in means ± SD and non-normal variables in medians and interquartile ranges. Bivariate analysis was performed using Fischer test, t test or Mann-Whitney U test. Multivariate analysis was carried out in a multiple logistic regression model. HR ApEn cut-off point was determined by the area under ROC curve. Significant p value was established < 0.05.

153 patients were studied. Mean age 42.5 ± 25 years. 86 males (56.2%). Sepsis was confirmed in 85 (55.5%). The overall median HR ApEn was 0.352. ApEn in septic patients was 0.414 vs 0.076 in non-septic patients, p < 0.001. ApEn cut-off point was 0.244 with a sensitivity of 97.7%, specificity of 90.6%, positive predictive value of 94.4% and a negative predictive value of 96.0%.

HR ApEn values greater than 0.244 predicts confirmed bacterial sepsis 24 hours before clinical presentation´s signs.