Markov model for detection of ECG instability prior to cardiac arrest in single-ventricle patients.

OBJECTIVES: Assess the degree of instability in the electrocardiogram (ECG) waveform in patients with single-ventricle physiology before a cardiac arrest and compare them with similar patients who did not experience a cardiac arrest. METHODS: Retrospective control study in patients with single-ventricle physiology who underwent Norwood, Blalock-Taussig shunt, pulmonary artery band, and aortic arch repair from 2013 to 2018. Electronic medical records were obtained for all included patients. For each subject, 6 h of ECG data were analyzed. In the arrest group, the end of the sixth hour coincides with the cardiac arrest. In the control group, the 6-h windows were randomly selected. We used a Markov chain framework and the likelihood ratio test to measure the degree of ECG instability and to classify the arrest and control groups. RESULTS: The study dataset consists of 38 cardiac arrest events and 67 control events. Our Markov model was able to classify the arrest and control groups based on the ECG instability with an ROC AUC of 82% at the hour preceding the cardiac arrests. CONCLUSION: We designed a method using the Markov chain framework to measure the level of instability in the beat-to-beat ECG morphology. Furthermore, we were able to show that the Markov model performed well to distinguish patients in the arrest group compared to the control group.

Quantification of electrocardiogram instability prior to cardiac arrest in patients with single-ventricle physiology.

OBJECTIVE: To quantify the instability measured in the electrocardiogram (ECG) waveform in patients with single-ventricle physiology before a cardiac arrest and compare with similar patients who did not have a cardiac arrest. METHODS: We measure the instability in the ECG morphology using variance, entropy, and decorrelation of polynomial fit coefficients of the beat-to-beat segmented data. These three metrics quantify the spread of the ECG morphology, the lack of beat-to-beat periodicity and the lack of predictability, respectively. For each subject, 3 h of ECG data were analyzed. In the arrest group, the end of the third hour coincides with the cardiac arrest. In the control group, the 3-h windows were randomly selected. RESULTS: The study dataset consists of 38 cardiac arrest events and 67 control events. In the hour prior to the cardiac arrest, the variance, entropy, and decorrelation of the polynomial fit coefficients were higher in the arrest group than in the control group (p = 0.003, p = 0.009, and p = 0.035, respectively). For the second and third hours prior to the arrests, the differences in variance, entropy, and decorrelation between the arrest and control groups lost statistical significance. Using these metrics of instability as predictive features in a support vector machine algorithm, we found an area under the receiver operating characteristic curve of 0.8 to distinguish the arrest event from the control events. CONCLUSION: By taking a holistic assessment of the ECG waveform in patients with single-ventricle physiology to measure the instability in its beat-to-beat morphology, the ECG waveform variance, entropy, and decorrelation are found to be statistically different in the patients who arrested compared with patients in the control group.