Multivariate Analyses with Power Spectral Measures of RR Variability. In multivariate Cox regression models using a step-up approach to evaluate the independent associations between frequency domain measures of RR variability and death of all causes, ULF power was selected first, i.e., had the strongest association with death. Adding VLF power or LF power to the Cox regression model significantly improved the prediction of death. With both ULF and VLF power in the Cox regression model, the addition of the other two components, LF and HF power, singly or together, did not significantly improve the prediction of all-cause mortality. Bigger et al. explored the relationship between the RR variability measures and all-cause mortality, cardiac death, and arrhythmic death before and after adjusting for five previously established postinfarction risk predictors: age. New York Heart Association functional class, rales in the coronary care unit, left ventricular ejection fraction, and ventricular arrhythmias detected in a 24-hour Halter ECG recording (38). Measures of RR variability, especially ULF and VLF power, were significantly and strongly associated with subsequent all-cause mortality and arrhythmic death after adjusting for the other risk predictors.
Improved Prediction of Death Using RR Variability Combined with Other Risk Predictors. Correlations between time or frequency domain measures of RR interval variability and previously identified post infarction risk predictors, e.g., left ventricular ejection fraction and ventricular arrhythmias, are remarkably weak (37). This study used the CHRONOS algorithms to calculate- time and frequency domain measures of RR interval variability. The strongest correlation is with RR interval itself (r=0.52). However, Fleiss et al. showed that adjusting for heart rate in various ways, e.g., coefficient of variation, did not improve the prediction of mortality (44). This lack of correlation suggests that measures of RR interval variability may improve the identification of patients at high risk after myocardial infarction. After adjustment for the five risk predictors, the association between mortality and total, ULF, and VLF power remained significant and strong, whereas LF and HF power were only moderately strongly associated with mortality. The tendency for VLF power to be more strongly associated with arrhythmic death than with all-cause or cardiac death was still evident after adjusting for the five covariates. Adding power spectral measures of RR variability, measured about 10 days after myocardial infarction, to previously known post infarction risk predictors identified small subgroups with a 2.5-year mortality risk of approximately 5070 (38).
RR Variability Index to Predict Arrhythmic Events Late after Myocardial infarction
Farrell et al. (45) studied the ability of RR variability to predict arrhythmic events (sustained ventricular arrhythmias or arrhythmic death) in 416 patients with acute myocardial infarction. The study group experienced 24 arrhythmic events and 47 deaths during follow-up. A 24-hour ECG and a signal averaged ECG were performed 6 or 7 days after myocardial infarction. RR variability was measured by triangular interpolation of the frequency distribution of RR intervals (the RR variability index). The mean value for RR variability index was 27ħl I msec (range 3 to 81 msec). Farrell et al. arbitrarily chose a value of 20 msec for RR variability index that dichotomized their group into 113 patients (27%) with low values and 348 patients (7370) with high values. The risk of experiencing an arrhythmic event during two years of follow-up was 32 times as high in the group with a RR variability index <20 msec as in the group with a value >20 msec. The risk of experiencing a cardiac death was about 7 times as high in the group with a RR variability index <20 msec. The RR variability index was a stronger univariate predictor of arrhythmic events or cardiac death than ventricular arrhythmias, signal averaged ECG, ejection fraction, exercise test, or coronary angiography. RR variability index did not predict recurrent myocardial infarction. The best multivariate predictor for arrhythmic events was the combination of RR variability index <20 msec, a positive signal averaged ECG, and repetitive ventricular premature complexes present in the same 24-hour ECG that was used to calculate RR variability. Figure 8 shows how combinations of risk predictors improve positive predictive accuracy or arrhythmic events during long-term follow-up after myocardial infarction. Adding the signal averaged ECG to the RR variability index doubled positive predictive accuracy, and adding repetitive ventricular premature complexes to the combination almost doubled positive predictive accuracy again. The combination of these three variables had a positive predictive accuracy of 5870.
The study of Farrell et al. suggests that measures of RR variability predict arrhythmic events better than non-arrhythmic deaths or recurrent ischemic events (45).
