Aim of the study was determination of physiological limits of QT-intervals and its derivative values in healthy children and adolescents during graded exercise tests. We examined 100 healthy boys and girls aged 11-15 years (mean age 13.4+/-2.1 years) and performed electrocardiography at rest and standard veloergometry (VEM) in all of them. We analyzed corrected intervals according to Bazett (QTc=QT/RR) and Fredericia (FQTc/3RR) formulas. Hysteresis QTc was calculated as difference between QTc durations during recovery and exercise at same heart rate (HR) Baseline HR before VEM exceeded rhythm on resting electrocardiogram by 5-15 bpm (84+/-8 vs 70+/-6, respectively, p<0.05) Increase of HR at exercise (mean 172+/-11 bpm) was similar in both sexes. QT interval decreased by 7-10% (18-31 ms) per each 25 w (p<0.05). Values obtained at determination of FQTc we found values 26-52 ms lower than those calculated by the Bazett formula in the process of whole test. Determination of FQTc compared with calculation by Bazett formula revealed more pronounced (10% from baseline level) shortening of FQT at peak exercise. QT calculated by the Bazett formula at 100 w did not differ from baseline level with tendency to higher level. Corrected QT according to the most often used Bazett formula was maximal at the first stage of exercise (25 w) and did not exceed 450 ms in boys and 460ms in girls. Maximal QTc lengthening in the process of test did not exceed 50 ms in any of the examined persons. Hysteresis of QTc interval was equal to 21+/-6 (15-25) ms. The conclusion was made that algorithm of assessment of QT interval changes during exercise test should include initial values of QTc calculated according to the Bazett formula, maximal QTc value, level of exercise at which it was registered, maximal increase of QTc during exercise, and QTc interval hysteresis.