Abstract 13937: Electrocardiography-Derived Respiratory Rate Detects Clinical Deterioration in Acute Care Patients With Cardiovascular Disease

• Published in: Circulation (New York, N.Y.) Vol. 134; no. Suppl_1 Suppl 1; p. A13937
• Main Authors: Moss, Travis J, Clark, Matthew T, Calland, James Forrest, Enfield, Kyle B, Kim, Gina, Lynch, Michelle A, Blatt, Marc I, Lake, Douglas E, Moorman, J Randall
• Format: Journal Article
• Language: English
• Published: by the American College of Cardiology Foundation and the American Heart Association, Inc 11.11.2016
• ISSN: 0009-7322; 1524-4539
• DOI: 10.1161/circ.134.suppl_1.13937

IntroductionClinical deterioration resulting in ICU transfer occurs in 4 to 5 of every 100 acute care admissions. Early Warning Scores (EWS) identify high risk patients, but rely on intermittent assessments. Continuous ECG not only provides heart rate (HR) and rhythm, but can also estimate respiratory rate (RR).HypothesisContinuous ECG data improves the predictive validity of EWS prior to acute deterioration leading to ICU transfer or unanticipated death.MethodsAmong floor patients admitted with continuous ECG, we identified those with deterioration resulting in ICU transfer or unanticipated death. We analyzed 60 patient-years of ECG data and applied previously validated methodologies for detection of atrial fibrillation (AF) and estimation of RR. We evaluated the predictive validity of common EWS and compared them to models that incorporate ECG data. We excluded observations after DNR/DNI orders or transitions to comfort care. We calculated the fold change in mortality, and the predictive validity (C-statistic).ResultsFrom 8,033 consecutive admissions, we identified 544 instances of clinical deterioration in 508 admissions. Admissions with events had a 50-fold increase in mortality (19.8% vs 0.4%) despite unexpected deaths accounting for only 6% of all events. We analyzed the 274 deteriorations that had ECG data in the 24-hours leading up to events. EWS had C-statistics ranging from 0.63 to 0.70. A model using only ECG-derived measurements had a C-statistic of 0.67 and the strongest predictors included RR, HR, and AF. Addition of the ECG-only model to the best EWS improved its C-statistic to 0.72.ConclusionsContinuous ECG data improves the ability of intermittent EWS to identify the highest risk patients up to 24-hours in advance of acute clinical deterioration.