A new method based on symbolic dynamics was applied to assess RR-QT dynamics and to compare gender differences. Segments of 10,000 RR and QT from the night were selected. The values of RR and QT were coded as follows. Each RR and QT interval was compared with their means in the last 50 beats [xRR, xQT]; when the interval was larger than x + delta then it was coded as a "2", where delta is the tolerance parameter; when it was less than x - delta-the code was a "0"; when it was larger than x-delta and and the less than x+delta-then it was coded as a "1." The tolerance parameter "delta" was equal to 10 ms for RR and 4 ms for QT. We obtained pairs of symbols representing the values of RR and QT-symbolic words. The results were presented in form of the probability density of the symbolic words. Mean RR, mean QT, SDRR, SDQT, QTc (Bazett formula) were also calculated. Electrocardiogram data of healthy individuals: 20 women and 20 men (mean age 39 +/- 12) were analyzed. There were significant gender differences in RR-QT dynamics. During heart rate acceleration the probability of QT shortening (the probability of the word "00") was higher in men than in women (P =.003). During heart rate deceleration QT lengthening (the word "22") was more frequently observed in men than in women (P =.003) as well. The QT reaction to RR interval changes is less complex in women than in men. In discriminant analysis, when QTc was ignored in the model, the RR-QT dynamics separated genders with 67% accuracy (chi(2) = 9.1, P <.003). RR-QT dynamics can be analyzed with symbolic dynamics methods. The gender differences in repolarization are not only due to QTc duration alone but also result from the dependence of the duration of QT on the RR duration.