Sinus node recovery time (SNRT) is frequently used to assess sinus node function in patients with suspected sick sinus syndrome (SSS). Although SNRT is assumed to reflect sinus node automaticity, this assumption remains unproven. The purpose of this study was (1) to test the hypothesis that SNRT in patients with and without SSS reflects sinus node automaticity, and (2) to assess the role of sinoatrial conduction time in the measurement of SNRT. A total of 16 patients (mean +/- SD age 63 +/- 9 years), seven of which had SSS, form the basis of this report. An electrogram of the sinus node was obtained for each of the 16 patients, and overdrive pacing was performed in each at cycle lengths of 1000 to 300 msec. SNRT was measured (1) on the sinus node electrogram (direct method, measuring SNRTd) as the interval from the last pacing stimulus artifact to the onset of the upstroke slope of first postpacing sinus beat and (2) on the high right atrial electrogram (indirect method, measuring SNRTi). Results were as follows: (1) The longest SNRTd was significantly shorter than the longest SNRTi (989 +/- 304 vs 1309 +/- 356 msec, p less than .001). (2) For the first postpacing sinus beat there was a significant prolongation of sinoatrial conduction time as compared with that for sinus beats before pacing (319 +/- 152 vs 99 +/- 35 msec, p less than .001). Sinoatrial conduction time normalized within 3.6 +/- 0.96 postpacing sinus beats. (3) At the pacing cycle length that resulted in the longest recovery time, sinus node depression was seen in 56% of patients, sinus node acceleration was noted in 26%, and no appreciable change in sinus node automaticity was observed in 19%. (4) Sinoatrial conduction time for the sinus beat before pacing and that for the first postpacing beat was longer in patients with SSS when compared with in patients without SSS. (5) In patients with SSS the abnormal SNRTi, when corrected for the degree of prolongation of sinoatrial conduction time for the first postpacing beat, became normal in five of six patients. We conclude that (1) SNRTi reflects both sinus node automaticity and sinoatrial conduction time, whereas SNRTd reflects sinus node automaticity, (2) overdrive atrial pacing results in marked prolongation of sinoatrial conduction time for the first postpacing beat, which is longer in patients with SSS when compared with in those without SSS, and (3) in patients with SSS the inference of abnormal sinus node automaticity on the basis of a prolonged corrected SNRTi is usually incorrect.