This study examines the accuracy of the conductance catheter technique and, in particular, parallel conductance [expressed as offset volume (Vc)] changes during the cardiac cycle in the human left ventricle. Two groups of patients were assessed: group 1, with an open atrial septal defect, and group 2, with an interventricular communication. In a subgroup, pre- and postoperative data were compared to assess the possible impact of shunting or anatomic considerations on our measurements. Vc is normally obtained by a saline-dilution technique previously described by Baan et al. [Vc(Baan); J. Baan, E. T. Van der velde, H. G. Debruin, G. J. Smeenk, J. Koops, A. D. Van Dijk, D. Temmerman, P. J. Senden, and B. Buis. Circulation 70: 812-823, 1984]. This does not take into account potential changes during the cardiac cycle. Four cardiac cycles were taken from the hypertonic saline washin and were divided into six equal isochrones between the maximum and minimum first derivatives of left ventricular pressure (dP/dtmax and dP/dtmin, respectively). The apparent ventricular volume was regressed against stroke volume for the corresponding cardiac cycle. The volume at the gamma-intercept corresponds to the Vc at each time interval [Vc(t)]. In group 1, there was a variation in Vc(t) during systole, but the temporal changes were quite small, on the order of 4.28% (SD = 5.18%) of total corrected end-diastolic volume (mean maximal variation of 2.60 ml). Furthermore, the value of Vc obtained at dP/dtmax was not significantly different from that obtained at dP/dtmin. For group 2 as a whole, mean Vc(Baan) did not change significantly with ventricular septal defect closure (preoperative, 8.85 +/- 11.1 ml; postoperative, 9.82 +/- 11.84 ml). Group 2 children also exhibited a systolic cyclical variation in Vc(t) similar to group 1. Finally, Vc(t) as a percentage of end-diastolic volume was no different when group 1 and group 2 were compared. We conclude that in the left ventricle, even in the presence of a left-to-right shunt, there is a small but insignificant difference in parallel conductance during ventricular ejection. The magnitude of this cyclical change does not preclude ventricular volume measurement in congenital heart disease by the conductance catheter technique.