Existing formulas for rate-corrected QT (QTc) commonly fail to properly adjust the upper normal limits which are more critical than the mean QTc for evaluation of prolonged QT. Age- and sex-related differences in QTc are also often overlooked. Our goal was to establish criteria for prolonged QTc using formulas that minimize QTc bias at the upper normal limits.Strict criteria were used in selecting a study group of 57,595 persons aged 5 to 89years (54% women) and to exclude electrocardiograms (ECG) with possible disease-associated changes. Two QT rate adjustment formulas were identified which both minimized rate-dependency in the 98th percentile limits: QTcmod, based on an electrophysiological model (QTcMod=QTx(120+HR)/180)), and QTcLogLin, a power function of the RR interval with exponents 0.37 for men and 0.38 for women. QTc shortened in men during adolescence and QTcMod became 13ms shorter than in women at age 20–29years. The sex difference was maintained through adulthood although decreasing with age. The criteria established for prolonged QTc were: Age <40years, men 430ms, women 440ms; Age 40 to 69, men 440ms, women 450ms; Age≥70years, men 455ms, and women 460ms.Sex difference in QTc originates from shortened QT in adolescent males. Upper normal limits for QTc vary substantially by age and sex, and it is essential to use age- and sex-specific criteria for evaluation of QT prolongation.