[New predictive methodology applied to the probabilistic loading of cardiac dynamics in a normal and chronic Holter]

To analyze cardiac dynamics by quantifying heart rate using equiprobable probability theory in order to discriminate between normal and chronic dynamics and to develop predictions of these dynamics.

Fifty 21-hour Holter recordings were evaluated. Based on two prototypes (normal and chronic diagnosis), repetitions of heart rates were taken in a probability space with ranges from 1001 to 2000 and from 2001 to 3000 repetitions. The probability of each repetition was calculated, and the most frequent probabilities in each range were summed. With the remaining tests, a blinded study was conducted, concealing the diagnoses and applying the same methodology; afterwards, the diagnoses were revealed and the probabilities were compared with those obtained in the initial induction. A blinded analysis was performed, calculating sensitivity, specificity, and kappa.

In normal dynamics, the sum of frequent probabilities ranged from 0.35 to 0.46 in the 1001-2000 range and from 0 to 1 in the 2001-3000 range; in chronic dynamics, from 0.35 to 0.48 and from 0.54 to 0.7, respectively. Normal frequencies were located between 55-95 and 65-95; in chronic dynamics, between 65-105 and 65-85. Sensitivity, specificity, and kappa were 1.

The probabilistic analysis of cardiac dynamics made it possible to define probability and frequency ranges that differentiate normal and chronic patterns, are predictive, and constitute an applicable diagnostic support tool.