OBJECTIVE--To investigate the relation between QT and RR intervals in the sick sinus syndrome or high degree atrioventricular block. PATIENTS--32 patients with episodes of prolonged RR intervals (> or = 2.6 s) on Holter electrocardiographic recordings. DESIGN--QT and RR intervals were measured manually every 100 to 150 beats on electrocardiographic strips reprinted from the Holter tape over 24 hours. The slope of the QT/RR relation was determined by the linear regression equation for RR intervals < or = 1.4 s (slope 1) and > 1.4 s (slope 2). RESULTS--Slope 2 (0.0068 (0.0030)) was significantly lower than slope 1 (0.0824 (0.0059), P < 0.0001) in the overall patient population. Slopes 1 and 2 were significantly lower (P < 0.001) in the 23 patients with QT intervals at the preceding RR interval of 1 s (QT1s) of < 0.44 s (0.0692 (0.0053) and 0.0019 (0.0030), respectively) than in the nine patients with QT1s intervals > or = 0.44 s (0.1159 (0.0091) and 0.0194 (0.0055), respectively). Slopes 1 and 2 correlated positively with QT1s interval in all patients. CONCLUSIONS--The QT/RR relation was comparatively flat when the RR interval was prolonged. Patients with prolonged QT intervals showed exaggerated prolongation of the QT interval with prolonged cycle lengths when compared with patients with normal QT intervals.