Atrioventricular Interval Optimization in the Right Atrial Appendage and Interatrial Septum Pacing: A Comparison Between Echo and Peak Endocardial Acceleration Measurements

PADELETTI, et al. : Atrioventricular Interval Optimization in the Right Atrial Appendage and Interatrial Septum Pacing: A Comparison Between ECHC and Peak Endocardial Acceleration. Interatrial septum pacing (IASP) reduces interatrial conduction time and consequently may interfere with atrioventricular delay (AVD) optimization. We studied 14 patients with an implanted BEST Living system device able to measure peak endocardial acceleration (PEA) signal. The aims of our study were to compare the (1) optimal AVD (OAVD) in right atrial appendage pacing (RAAP) and IASP, and (2) OAVD derived by the PEA signal versus OAVD derived by Echo/Doppler evaluation of the left ventricular filling time (LVFT) and cardiac output (CO). Measurements were performed in DDD VDD modes Eight patients (group A) had RAAP and six patients (group B) had IASP. In group A, OAVD measured by LVFT, CO, and PEA was 185 ± 23 ms , 177 ± 19 ms , and 192 ± 23 ms in DDD and 147 ± 19 ms , 135 ± 27 ms , and 146 ± 20 ms in VDD, respectively. OAVD measured by LVFT, CO, and PEA was significantly longer in DDD mode than in VDD (P < 0.01, P < 0.01, P < 0.001 ). In group B, OAVD measured by LVFT, CO, and PEA was 116 ± 19 ms , 113 ± 10 ms , and 130 ± 30ms in DDD and 106 ± 16 ms , 96 ± 15 ms , and 108 ± 26 ms in VDD, respectively. No statistical differences were observed between DDD and VDD. Significant correlations between OAVDs PEA derived and OAVDs LVFT and CO derived were observed (r = 0.71, r = 0.69, respectively ). When new techniques of atrial stimulation, as IASP, are used an OAVD shorter and similar in VDD and DDD has to be considered. The BEST Living system could provide a valid method to ensure, in every moment, the exact required OAVD to maximize atrial contribution to CO.     

Exercise Induced Sympathetic Influences Do Not Change Interatrial Conduction Times in VDD and DDD Pacing

Using telemetry, right atrial electrogram (RA), and marker channel of atrial sense events (M_A) in combination with the left atrial electrogram (LA), recorded by a filtered bipolar esophageal lead, interatrial conduction during submaximal exercise and at rest was examined in 46 DDD pacemaker patients. The RA-LA and M_A-LA conduction times measured in the presence of atrial sensing (VDD) as well as the conduction time S_A-LA from atrial stimulus (S_A) to LA, determined during atrial pacing (DDD) were found to be individual constants independent of exercise induced sympathetic influences. Thus, having determined an optima! mechanical interval (LA-LV)_mech/opt from left atrium to ventricle by other methods, the optimal AV delay for DDD as well as for VDD operation can be calculated by the sum of the appropriate interatrial conduction time (S_A-LA, respectively M_A-LA) and the (LA-LV)_mech/opt interval. Due to the constant S_A-LA and M_A-LA, the difference between these two values (AV delay correction interval) is a constant as well, which remains unchanged during exercise. Therefore, in selecting the rate responsive AV delay, only hemodynamic and not electrophysiologica] measurements need to be considered.     

Prediction of interatrial and interventricular electromechanical delays from P/QRS measurements: value for pacemaker timing optimization

Summary: Cardiac pacing creates spurious delays between and within the cardiac chambers. These are: 1. Left atrial (LA) transport delay (ATD) either sensed (s) or paced (p), (time from right atrial P-wave to the end of LA transport (mitral Doppler A-wave)). 2. Interventricular delay (IVD), (time from onsets of right (RV) to left ventricular (LV) contractions). 3. P-sense offset (PSO), (time from P-onset to P-detection). Thus, restoration of left heart atrioventricular (AV) synchrony can be accomplished by compensating above delays, according to a previously published equation: RAV = ATD-IVD-PSO, where RAV = right heart AV.
Objective: To test the hypothesis that ATD could be predicted from Ps and Pp, and that interventricular delay (IVD) could be predicted from QRSp, using three-lead surface electrocardiograms (ECGs).
Methods: Thirty-six patients aged 63.5 ± 15.5 years, 64% males, all with previously implanted DDD pacemakers, were studied by echo-Doppler and surface ECG obtained with a pacemaker programmer. Measurements included Ps; Pp; intrinsic QRSs; and paced QRSp, ATDs, ATDp, and IVD (difference between RVp and RVs left preejection intervals, PEI). Regressions between ECG and echo-Doppler intervals were calculated.
Results: Regressions and correlation coefficients: ATD (s+p) = 0.96*P + 55 (R = 0.94, P < 0.0001); PEIp = 0.75 * QRSp + 34.8 (R = 0.89, P< 0.0001); IVD = 0.39 *QRSp – 7.9 ms (R = 0.87, P = 0.002).
Conclusions: Inter-atrial and inter-ventricular electromechanical delays can be predicted from P-wave and QRS durations. These measurements allow AV delay optimization in DDD and cardiac resynchronization therapy devices with no need of Doppler echocardiography.     

The Importance of Different Atrioventricular Delay for Left Ventricular Filling in Sequential Pacing: Clinical Implications

We assessed the influence and clinical consequences of different AV delay on ventricular filling in 30 patients (mean age 60 ± 5 years) who had ODD pacemakers for AV block. All 30 patients presented a normal ejection fraction, but in 18 cases (Group I), an echo-Doppler examination revealed ventricular hypertrophy (mean end-diastolic wall thickness of 1.4 ± 0.16 cm. LV mass index 155 ± 17g/m²), and an abnormal relaxation pattern (isovoiumetric relaxation time = 124.72 ± 11.82; early to late peak velocity = 0.6 ± 0.03; deceleration time - 296.83 ± 34.02 ms). Group II included the remaining 12 patients who had a normal filling pattern. In all 30 patients, the pattern was reassessed following modification of the AV delay (200, 150, 100. and 75 ms). Patients at baseline (AV delay of 200 ms) also underwent an exercise test with determination of respiratory gas exchange. In Group I, 13 (72.5%) patients were classified as Weber class B (VO₂ Max 16.8 ± 1.7mL/min per kg); and 5 (27.5%) were Class A (VO₂ Max 22.5 ± 1.4 mL/min per kg). In Group II, all 12 patients were classified as Weber Class A. In Group II, changes in AV delay caused no consistent variations in filling pattern, and therefore AV delay was not modified. In Group I patients, since reduction to 100 ms resulted in normalization of the filling pattern, the AV delay was programmed to 100 ms. A graded exercise test repeated after 6 months follow-up showed an improved Weber class in 13 patients (from B to A) and greater VO₂ Max in the remaining five already in Class A. We concluded that, in sequential paced patients with normal ejection fraction but abnormal relaxation pattern, modification in AV delay can induce normalization of filling and improvement in cardiac functional capacity.     

Effects of Different Pacing Modes on Left Ventricular Relaxation in Closed-Chest Dogs

Ventricular relaxation is an important determinant of ventricular filling; impaired relaxation may decrease cardiac output and stroke volume. Relaxation hos been shown to occur more quickly following beats with an increased extent of systolic fiber shortening. Since cardiac output and stroke volume are greater during atrioventricular (AV) sequential pacing than during ventricular pacing at identical heart rates, we reasoned that AV sequential pacing would improve relaxation. To assess this hypothesis we studied 11 dogs with chronic (1-3 months) complete heart block (CHB) induced by radiofrequency catheter ablation of the His bundle. Right and left heart pressures, thermodilution cardiac output, und single plane ventriculography were recorded during baseline rhythm (CHB), and VVI, and AV sequential pacing at a heart rate greater than the sinus rate. None had ventriculoatrial conduction. During AV sequential pacing, the AV interval was set at 150 msec. Cardiac output and stroke volume were significantly increased in the AV sequential compared to the VVI pacing mode. Left ventricular pressures, maximal positive and negative dP/dt, and the time constant (T) of isovolumic pressure decay were not different in the two modes. We conclude that despite increased stroke volume in the AV sequential pacing mode, relaxation is unchanged. We believe the lack of change in relaxation is due to nonuniform ventricular activation when depolarization is initiated at the right ventricular apex.     

Electrocardiographic Characteristics of Pacing From the Right Atrial Appendage During Atrioventricular Sequential Pacing

In the interpretation of electrocardiograms recorded during atrioventricular sequential pacing, uncertainty frequently arises in the assessment for evidence of atrial capture. In the present study, electrocardiographic characterization of pacing from the right atrial appendage as a component of atrioventricular pacing was performed on tracings obtained from 16 patients with bipolar dual chamber pacing units, and from 18 patients with unipolar dual chamber pacing units in which large overshoot potentials occurred following the atrial pacing spike. Atrial complexes resulting from bipolar pacing of the right atrial appendage were found to be uniformly prolonged and of diminished amplitude compared to those in sinus rhythm; they were also noted to contain sequential inferoposterior and leftward-posterior component vectors. The exponential overshoot-decay complex associated with unipolar atrial pacing appeared as a vector directed along the axis from the pulse generator to the pacing lead; the degree to which this deflection interfered with identification of atrial capture in various leads was thus largely dependent on pulse generator location. It was concluded that careful systematic inspection of multiple electrocardiographic leads will generally permit the characteristic features of pacing the right atrial appendage to be recognized, thus facilitating correct interpretation of atrial capture during atrioventricular sequential pacing.     

A Time-Related Study of the Hemodynamic Benefit of Atrioventricular Synchronous Pacing Evaluated by Doppler Echocardiography

We have used Doppler echocardiography to estimate the stroke volume (SV) in a study of 13 patients equipped with DDD pacemakers. SV was measured both during DDD and VVI pacing after observation times of 1,3,6, and 12 months of DDD pacing. SV was also measured at seven atrioventricular (AV) intervals (75-250 ms) in the search for optimal AV intervals. Mitral flow velocity was investigated to see if DDD pacing resulted in synchronous atrial contraction, and if mitral insufficiency existed at any of the pacing modes. Compared with the VVI mode, DDD pacing resulted in a mean increase in SV of 21 +/- 2% for the four observation periods. Two patients with severe left ventricular failure had no significant increase in SV during DDD vs VVI pacing. In each patient, an optimal AV interval ranging between 100-250 ms for the SV was found. Velocity profiles of mitral flow showed synchronous atrial contraction during DDD pacing, but not during VVI pacing. Mitral insufficiency was not seen in any pacing mode. DDD pacing resulted in a reduction in SV during the first 6 months, and was constant thereafter. Doppler echocardiography can be used repeatedly to evaluate the hemodynamic response of DDD pacing vs VVI pacing, and to find which AV interval gives the highest SV in the individual patient. Our study further shows that the hemodynamic benefit of DDD pacing is present after short-term as well as after long-term DDD pacing.     

Doppler Echocardiographic Studies in Sequential Atrioventricular Pacing

Six patients with sequential atrioventricular pacemakers were studied by Doppler echocardiography. A commercially available continuous wave system (transcutaneous aortovelography) was used and the transducer was placed in the suprasternal notch and angled to obtain peak aortic arch blood flow velocity. Changes in Doppler peak velocity have been previously shown to correlate closely and reliably with changes in hemodynamically measured stroke volume/cardiac output in the same patients following interventions, In all patients, the pacemaker was programmed from DVI mode to VVI (heart rate was kept constant) and then back to DVI, with the Doppler transducer held in a fixed position all the time, and peak velocities measured after the patient had been in any particular mode for at least 1 minute. Five of 6 patients showed reduction in peak velocities ranging from 13–25% (mean 18.4%) when the pacing mode was switched from DVI to VVI. Doppler peak velocity remained unchanged in 1 patient. Doppler echocardiography represents a simple, noninvasive method for estimating increments in stroke volume/cardiac output obtained with sequential AV pacing.     

Effect of Atrioventricular Sequential Pacing on Left Ventricular Flow Dynamics in a Patient with Mid-Ventricular Obstruction

The effect of dual chamber atrioventricular sequential pacing on the intraventricular pressure gradient was tested using Doppler echocardiography in a patient with hypertrophic mid-ventricular obstruction. Atrioventricular sequential pacing with relatively short atrioventricular delays reduced in the left ventricular pressure gradient at the mid-ventricular level. Also, atrioventricular sequential pacing affected the degree and profile of the isovolumetric relaxation flow. These results suggest that atrioventricular sequential pacing affects both systolic and diastolic left ventricular flow dynamics in mid-ventricular obstruction.     

Relationship Between Interatrial Conduction Times and Left Atrial Dimension in Patients Undergoing Atrioventricular Stimulation

Interatrial conduction time (IACT) and left atrial dimension (LAD) were determined in 75 patients (41 males, 34 females, mean age 78.2 ± 7,9 years) undergoing atrioventricular (AV) stimulation. The LAD was measured by M mode echocardiography as the distance between the posterior aortic wall and the posterior left atrial wall. The IACT was determined during a transvenous dual chamber pacemaker implant done under local anesthesia (lidocaine). The spontaneous interatrial conduction time (SIACT) was measured from the intrinsic deflection (ID) of the right atrium recorded in a unipolar mode (unipolar J-shaped had positioned in the right appendage) to the ID of the left atrium (bipolar esophageal lead, left atrial positive deflection equal to the negative one) during sinus rhythm. The right atrium then was paced at a rate slightly greater than the spontaneous one. The paced interatrial conduction time (PIACT) was measured from the stimulus artifact to the left atrial ID. The PIACT was also measured during incremental right atrial pacing (10 beats/min step increase to 180 beats/min) and, from these measurements, the maximum increase of PIACT (MIPIACT) was deduced. The LAD was measured at 39.5 ± 8.7 mm, SIACT at 70.3 ± 24.8 msec, PIACT at 118.8 ± 27.9 msec, and MIPIACT at 16.5 ± 16.4 msec. We found highly significant relationships between SIACT and LAD(P = 0.0006, r - 0.39), PIACT and LAD (P = 0.0001, r = 0.45), and MIPIACT and LAD (P = 0.0006, r = 0.38). We also noted that the LAD was greater in patients in whom MIPIACT was >10 msec than in patients in whom the MIPIACT was negligible (P < 0.002). However, the “r” values indicate that IACT is probably determined by multiple factors, and LAD appears to be one of the most important. Thus, we demonstrated the existence of highly significant relationships between the LAD determined by M mode echocardiography and the IACT when sensing and pacing the right atrium. We also demonstrated that the LAD was greater in patients in whom PIACT increased by an appreciable duration during fast atrial pacing. These results must be kept in mind when choosing a mode of stimulation and determining the AV delay (dual chamber pacemaker), particularly in patients with left atrial enlargement in whom the contribution of the atrial contraction and its timing are hemodynamically determinant.     

Lack of Influence of Atrioventricular Delay on Stroke Volume at Rest in Patients with Complete Atrioventricular Block and Dual Chamber Pacing

Dual chamber pacing (DDD) maintains atrioventricular (AV) sequence; AV delay programmability modifies the relationship between atrial and ventricular contraction. To evaluate the hemodynamic effects of such a modification, ten patients with a DDD unit for complete AV block were studied by time-motion (M-mode) and Doppler echocardiography during inhibited ventricular pacing (VVI), atrial-triggered ventricular pacing (VDD) and atrioventricular sequential pacing (DVI) at different AV delay (90, 140, 190, 240 msec). A significant improvement in stroke volume (SV) (15%-20%, P less than 0.05) was seen during DDD versus VVI pacing; no changes, however, were observed in the same patient with different AV delay or during DVI versus VDD pacing. These data suggest that programming of AV delay does not affect systolic performance at rest; longer diastolic filling times recorded during DDD pacing with "short" AV delay (90-140 msec) do not seem to be a hemodynamically relevant epi-phenomenon of PM programming.     

Assessment of Atrial Capture in Committed Atrioventricular Sequential (DVI) Pacing Systems

Two patients with committed dual unipolar AV sequential pacing systems are presented in whom alrial capture was unable to be demonstrated on the surface electrocardiogram. The bedside examination is proposed as a means of identifying the presence or absence of atrial capture with subsequent confirmation by M-mode echocardiographs and jugular venous pulse tracings. Mode, rate, and output programmabilitymayaidin the evaluation of thesepacing systems. Intentional reduction of the ventricular output to subthreshold levels in pacemakers capable of this degree of programmabilily will permit the rapid bedside confirmation of atrial capture without the need for expensive and time-consuming noninvasive tests.     

The effect of the atrial pacing site on the total atrial activation time

The effect of dual site pacing for prevention of atrial fibrillation may be due to synchronization of right and left atrial activation. Little is known, however, about the effect of pacing from single right atrial sites on differences in interatrial conduction. Twenty-eight patients without structural heart disease were studied following radiofrequency catheter ablation of supraventricular arrhythmias. Pacing was performed using standard multipolar catheters from the presumed insertion site of Bachmann's bundle, the coronary sinus ostium, the high lateral right atrium, and the right atrial appendage (n = 8 patients). Bipolar recording was performed from the distal coronary sinus, the high and low lateral right atrium, and the posterolateral left atrium (n = 13 patients). The longest conduction time from each pacing to each recording site was considered the total atrial activation time for the respective pacing site. During high right atrial pacing, the total atrial activation time was determined by the conduction to the distal coronary sinus (118 +/- 18 ms), during coronary sinus ostium pacing by the conduction to the high right atrium (94 +/- 18 ms), and during Bachmann's bundle pacing by the conduction to the distal coronary sinus (74 +/- 18 ms). The total atrial activation time was significantly shorter during pacing from Bachmann's bundle, as compared to pacing from other right atrial sites. Thus, in normal atria, pacing from the insertion of Bachmann's bundle causes a shorter total atrial activation time and less interatrial conduction delay, as compared to pacing from other right atrial sites. These findings may have implications for alternative pacing sites for prevention of atrial fibrillation.     

Impedance Cardiography for Atrioventricular Interval Optimization During Permanent Left Ventricular Pacing

TSE, H-F., et al. : Impedance Cardiography for Atrioventricular Interval Optimization During Permanent Left Ventricular Pacing. Left ventricular (LV) pacing is increasingly used in the management of congestive heart failure. Optimization of the atrioventricular (AV) interval is essential to maximize the hemodynamic benefits of this therapy. Although Doppler echocardiography (echo) is the most widely used method, it is time-consuming, expensive, and operator-dependent. We examined the value of an impedance cardiography (IC)-based method of cardiac output (CO) measurement to optimize the AV interval in 5 men and 1 woman (mean age   = 72 ± 11   years) during permanent LV pacing with a 4.8 Fr unipolar coronary sinus pacing lead. Simultaneous measurements of CO by IC and echo were performed at AV intervals of 50, 80, 110, 150, 180, and 225 ms during DDD pacing at 85 beats/min. The optimal AV interval varied between 110 and 180 ms. In 5 of 6 patients (83%), the optimal AV interval by echo and IC was identical. While CO measurements were higher with IC than with echo (   6.1 ± 0.4 L/min   vs 4.7 ± 0.3 L/min, P < 0.05), CO measurements by IC and echo were closely correlated   r = 0.67   , P < 0.001). In conclusion, our initial experience suggests that IC is a reliable method of AV interval optimization during LV pacing. IC and echo measurements of CO during LV pacing were closely correlated. (PACE 2003; 26[Pt. II]:189–191)     

Is Vagal Innervation to the Atrioventricular Node Impaired After Radiofrequency Ablation of the Slow Atrioventricular Nodal Pathway?

To assess the potentially adverse effects of RF catheter ablation (RFCA) of the slow AV nodal pathway on the parasympathetic innervation to the AV node in patients with AV nodal reentrant tachycardia (AVNRT), AV nodal conduction was evaluated following vagal stimulation by means of a phenylephrine bolus injection (200 μg) before and after RFCA in ten patients (mean age, 37 ± 14 years). Nine patients with AV reentrant tachycardia (AVRT) due to a left free wall accessory pathway served as a control group (mean age of 37 ± 12 years). Whereas no prolongation of the AH interval was observed in the AVNRT group following the phenylephrine bolus during sinus rhythm, despite a significant slowing in sinus rate, phenylephrine administration in AVRT patients was associated with both slowing of the sinus rate and prolongation of the AH interval. Following successful RFCA, the same responses were observed. To delineate the indirect effect of heart rate on AV conduction in response to the phenylephrine bolus, the AH interval was also measured during fixed atrial pacing. A marked prolongation of the AH interval occurred in both groups following phenylephrine administration. This prolongation was biphasic in 50% of A VNRT patients before ablation, suggesting a predominant effect of vagal stimulation on the fast AV nodal pathway. RFCA was associated with disappearance of discontinuous AV conduction in all but one patient with AVNRT. Vagal stimulation caused the same amount of AH interval prolongation as before RFCA in both study groups. In conclusion, patients with AVNRT have a preserved modulation of AV nodal conduction in response to vagal stimulation during sinus rhythm. In addition, vagal stimulation seems to exert a predominant effect on the fast A V nodal pathway. RFCA of the slow AV nodal pathway in patients with A VNRT does not cause detectable damage to the vagal innervation to the AV node.     

Atrioventricular Sequential Pacing: Comparison with Ventricular Pacing Using Systolic Time Intervals

Nine patients with programmable atrioventricular sequential pacers were studied using systolic time intervals (QS2—the total electrical and mechanical systole, left ventricular ejection time, and pre-ejeclion period). These measurements were obtained by simultaneous recording of the electrocardiogram, phonocardiogram. and carotid pulse tracing. There was a dramatic fall of left ventricular ejection time (LVET) and an increase of the pre-ejection time (PEP) in all patients when the pacers were programmed from the atrioventricular to the ventricular mode at constant heart rate. This resulted in an increase of the ratio PEP/LVET from 428 to 574 suggesting loss of ventricular function. These measurements all reversed to baseline values when the pacers were reprogrammed back to the atrioventricular mode. This study suggests systolic time intervals might be useful to select non-invasively pacer parameters such as mode, rate, and effective PR interval in order to provide the best hemodynamics in a given patient.     

Programming a Long Paced Atrioventricular Interval May Be Risky in DDDR Pacing

In patients with intermittent AV block and dual chamber pacemakers, a long paced AV interval of 200 msec or more can be selected to prolong pulse generator life (by avoiding the ventricular pace output) and to enable a more physiological and hemodynamically superior activation sequence. This case report describes the potential risks of programming a long paced AV interval in a patient with a DDDR pacemaker. T wave pacing, as described here, can occur if the conducted QRS complex is not sensed because it occurs during the ventricular blanking period (delivery of the atrial stimulus). This can be initiated by the mechanisms that induce apparent and actual P wave undersensing of the conducted QRS complex. In this case report apparent P wave undersensing and subsequent T wave pacing with ventricular capture (in a patient with intermittent AV block) occurred frequently during an exercise test done in the DDDR mode with a paced AV interval of 200 msec, according to the clinical evaluation protocol.     

Dual Chamber Pacing in Hypertrophic Obstructive Cardiomyopathy: Beneficial Effect of Atrioventricular Junction Ablation for Optimal Left Ventricular Capture and Filling

Clinical improvement with dual chamber pacing bas largely been reported in patients suffering from hypertrophic obstructive cardiomyopathy and mainly attributed to the reduction of the subaortic pressure gradient. To be effective, pacing must induce a permanent and complete capture of the LV. In two patients of our collective, symptoms (angina and dyspnea NYHA Class III and/or syncopes) persisted or relapsed despite pacing. This was related to the inability to obtain full LV capture due to a too-short native PR interval. RF ablation of the AV junction was therefore performed in botb patients, resulting in permanent AV block in one and prolonged PR interval up to 310 ms in the second. Pacing was thereafter associated with an immediate and significant clinical improvement related to permanent LV capture, whatever the patient's activity. After RF ablation, the AV delay was set up to induce the best LV filling, as assessed by Doppler analysis of mitral flow. Our observations suggest that RF ablation or modification of the AV junction can be a successful procedure in some patients with residual or recurrent symptoms, when the latter result from a loss of capture or from the inability to program an AV delay tbat does not compromise the active component to LV filling. Doppler echocardiography is a simple and effective mean to assess the hemodynamic effect of AV interval modulation in this setting.     

Atrioventricular Nodal Conduction and Refractoriness Following Abrupt Changes in Cycle Length

The properties of the atrioventricular (AV) nodal conduction and effective refractory period in man are generally evaluated at a constant basic cycle length (CL) and, in most cases, they demonstrate an inverse relationship to the drive cycle. The response of AV node to abrupt change in CL is less defined. We therefore studied the effects of abrupt changes in CL on AV nodal conduction time and refractoriness in 18 patients. AV nodal conduction time, and effective and functional refractory periods were measured during: (1) a constant long CL, (2) a constant short CL, and (3) after an abrupt increase in CL just prior to the introduction of extrastimuli. In 10 of the 18 patients a constant long CL of 600 ms, a constant short CL of 400 ms and a sudden short-to-long change in CL (400 to 600 ms) were tested. AV nodal conduction times (A2H2) were measured at the shortest and longest comparable A1A2 intervals. The mean value of the shortest A2H2 intervals for constant CL of 600 ms was 144 +/- 18 ms; for a constant CL of 400 ms it was 162 +/- 17 ms; after a sudden short-to-long change in CL (400 to 600 ms) it was 142 +/- 14 ms. The mean value of the longest A2H2 intervals at a constant CL of 600 ms was 185 +/- 18 ms; at a constant CL of 400 ms it was 236 +/- 26 ms (p less than 0.01) and after a short-to-long change in CL (400 to 600 ms) 199 +/- 21 ms. AV nodal effective refractory periods measured at the same three CLs had mean values of 279 +/- 13 ms; 300 +/- 15 ms and 294 +/- 13 ms, respectively. Similar results were obtained when other CLs such as 700 to 900, 500 to 900, and 400 to 700 ms were tested. The data suggest that after abrupt short-to-long changes in CL, AV nodal function curves shift from long constant CL toward short constant CL as the coupling intervals decrease, indicating a cumulative pattern. Although the return to baseline conduction time after the fast basic rate is known to be slow, the limitation of this effect to the very early premature beat in the human has not been reported previously.