Background The diagnosis of the long-QT syndrome (LQTS) may be difficu
lt to establish in patients with normal or borderline prolongation of
the QT interval. Noninvasive markers are needed to identify patients w
ith LQTS. Methods and Results Fourteen patients with known LQTS, 9 una
ffected family members, and 40 control subjects underwent modified Bru
ce protocol exercise testing. The RT interval (peak of R wave to peak
of T wave) and rate-corrected RT interval (RTc) were measured during e
xercise and recovery. The RT interval at 1 minute into recovery was su
btracted from the RT interval at a similar heart rate during exercise
(ART). The RTc shortened by 61 milliseconds (ms) in the LQTS patients
compared with 23 to 26 ms in the other two groups (P=.003 by ANOVA). T
he RT interval shortened in a linear fashion in all patients but demon
strated persistent shortening during recovery in the LQTS patients. Th
is was manifested as a hysteresis loop in the curve relating the RT in
terval to cycle length. The hysteresis loop was present in 13 of 14 LQ
TS patients and only 4 of 40 control subjects. Delta RT >25 ms had a s
ensitivity of 73%, a specificity of 92%, a positive predictive value o
f 79%, and a negative predictive value of 90% for LQTS. Conclusions Hy
steresis of the RT interval with exercise may be useful for the diagno
sis of LQTS.