Assessment of mean respiratory rate from ECG recordings for risk stratification after myocardial infarction.

ABSTRACT

BACKGROUND: We recently reported that nocturnal respiratory rate (NRR) predicts non-sudden cardiac death in survivors of myocardial infarction (MI). Here, we present the details of the technique deriving NRR from ECG recordings. METHODS: Continuous ECG and respiratory chest excursions were simultaneously recorded in 941 MI survivors who were followed-up for 5-years. Mean respiratory rate was derived from the ECG based on RR intervals, QRS amplitudes, and QRS vectors and compared to chest belt measurements. NRR was calculated from Holter-ECGs accordingly using the same ECG processing. RESULTS: Directly-measured and ECG-derived respiratory rates were in good agreement. Areas under the ROC curve for 10-min-ECG- and Holter-derived respiratory rate were well in the confidence intervals of that of the chest belt measurement. The optimum dichotomy of NRR for the prediction of mortality was ≥18.6 breaths per minute. CONCLUSIONS: The mean respiratory rate can be precisely derived from continuous ECGs.