Aarhus University Seal / Aarhus Universitets segl

Johannes Enevoldsen

Prevalence and Temporal Distribution of Extrasystoles in Septic ICU Patients: The Feasibility of Predicting Fluid Responsiveness Using Extrasystoles

Research output: Contribution to journal/Conference contribution in journal/Contribution to newspaperJournal articleResearchpeer-review



Extrasystoles may be useful for predicting the response to fluid therapy in hemodynamically unstable patients but their prevalence is unknown. The aim of this study was to estimate the availability of extrasystoles in intensive care unit patients diagnosed with sepsis. The study aim was not to validate the fluid responsiveness prediction ability of extrasystoles.

Twenty-four-hour ECG recordings from a convenience sample of 50 patients diagnosed with sepsis were extracted from the MIMIC-II waveform database, and ECGs were visually examined for correct QRS complex detection. Custom-made algorithms identified potential extrasystoles based on RR intervals. Two raters visually confirmed or rejected the potential extrasystoles and then classified them as ventricular, supraventricular, or unknown origin. Extrasystole availability was calculated as extrasystolic coverage for each 24 h ECG recording, that is, the percentage of the 24 h recording where an extrasystole had occurred in the preceding 30 minutes.

Mean extrasystolic coverage was 53.3% (confidence interval: [42.8; 63.6]%) and ventricular extrasystolic coverage was 21.4 [13.5; 29.8]%. Interrater reliability was strong for confirming/rejecting extrasystoles.

Extrasystoles are available for fluid responsiveness prediction in septic patients in about half of the time. With this extrasystolic availability, we believe the method to be considered for clinical use, provided that future studies validate the method's fluid responsiveness prediction ability.
Translated title of the contributionForekomst og tidsmæssig fordeling af ekstrasustoler hos septiske intensivpatienter: Anvendeligheden af ekstrasystoler til at forudsige væskerespons
Original languageEnglish
Article number5697092
JournalCritical Care Research and Practice
Pages (from-to)1-8
Number of pages8
Publication statusPublished - 1 Oct 2018

See relations at Aarhus University Citationformats

ID: 136064949