Author + information
- Received January 14, 2000
- Revision received April 27, 2000
- Accepted June 21, 2000
- Published online November 1, 2000.
- ↵*Reprint requests and correspondence: Dr. Brian D. Hoit, Division of Cardiology, Case Western Reserve University, 11100 Euclid Avenue, Cleveland, Ohio 44106.
The purpose of this study was to understand the significance of an effective atrial systole and the interactions between atrial and ventricular function.
The significance of atrial function is controversial, particularly in the setting of left ventricular (LV) dysfunction.
Serial, rapid pacing in five dogs that had undergone radiofrequency ablation and implantation of right atrial and ventricular pacemakers produced reversible atrial and ventricular dysfunction (alone and in combination). Atrial function (echocardiograph-determined transmitral diastolic flow, left atrial appendage emptying, and pulmonary venous flow), cardiac output, and right heart pressures were measured at matched paced heart rates of 80 beats/min.
Isolated rapid atrial pacing (LV ejection fraction ∼60%) decreased atrial booster pump in the body and appendage of the left atrium, but increased the conduit function of the left atrium. Isolated LV dysfunction (LV ejection fraction ∼34%) increased atrial booster pump function. The decreased atrial booster pump function in animals with combined atrial and ventricular dysfunction was incompletely compensated by the redistribution of the reservoir and conduit functions of the left atrium. As a result, cardiac output decreased and right heart pressures increased only after superimposed pacing.
In the presence of a normal left ventricle (LV), atrial failure has little effect on cardiac output and right heart pressures because of compensatory conduit function, but when early LV dysfunction coexists, changes in reservoir and conduit functions are insufficient to compensate for an impairment of atrial contraction.
The principal function of the left atrium (LA) is to modulate left ventricular (LV) filling and cardiovascular performance. This is accomplished through its well-defined actions as a booster pump during late ventricular diastole, and as a distensible reservoir and passive conduit for pulmonary venous flow during ventricular systole and early ventricular diastole, respectively (1–6). However, the significance of atrial function remains arguable, particularly in the context of abnormal LV function (7–9). Limitations of experimental designs and analytic techniques and the variable and often unknown degrees of concomitant atrial and ventricular dysfunction are largely responsible for this controversy.
A reciprocal relation exists between conduit and reservoir functions of the LA. Indeed, the redistribution of these atrial functions was reported as an important compensatory mechanism that facilitates LV filling in patients with myocardial ischemia (10), acute myocardial infarction (11), and hypertensive heart disease (12). However, little is known how these mechanisms operate in the setting of LV dysfunction. Accordingly, we undertook the present study to test the hypothesis that, in contrast to dogs with a normal LV, cardiac output falls with the loss of atrial systolic contraction in dogs with concomitant LV dysfunction because atrial compensatory mechanisms fail.
To accomplish these aims, chronic atrial and ventricular dysfunction (both singly and in combination) were modeled with rapid pacing (13). Serial echocardiographic analyses of global and regional LA pump function were estimated from mitral and LA appendage Doppler velocimetry, and relative reservoir and conduit functions of the LA were determined from pulmonary venous Doppler (1). Right heart catheterization studies were performed simultaneously to assess right heart pressures and cardiac output.
Six mongrel dogs weighing 23 to 30 kg were prepared for serial echocardiographic-hemodynamic studies with radiofrequency (RF) atrioventricular (AV) nodal ablation and implantation of right atrial and ventricular pacing catheters. Complete studies were obtained in five of the animals, which form the basis of this report. This animal investigation conforms to the guideline “Position of the American Heart Association on Research Animal Use.”
Isolated atrial systolic failure was produced by one week of rapid atrial pacing (400 beats/min), and moderate LV dysfunction was produced by two weeks of rapid right ventricular pacing (220 beats/min). None of the animals developed physical signs of heart failure. Rapid pacing of the right ventricle for two weeks, and superimposing rapid atrial pacing during the second week, produced combined LA and LV dysfunction. The sequence of animal modeling was random. After the animal recovered (one week after cessation of atrial pacing, two weeks after cessation of ventricular pacing) the pacemakers were reprogrammed to produce the next model. Doppler echocardiographic studies in three of the animals confirmed that atrial and ventricular function returned to normal one to two weeks after pacing was stopped (data not shown).
Radiofrequency AV nodal ablation and RV pacemaker implantation
Animals were anesthetized with morphine sulfate (3 mg/kg) and pentobarbital (20 mg/kg), secured on a fluoroscopy table, and intubated. A 7F transvenous RF ablation electrode was advanced from the femoral vein to the right atrium and was connected to a Radionics Model RFG-3 graphic RF Lesion Generator System (Radionics, Burlington, Massachusetts) interfaced to a Gould Windowgraf 4-channel recorder (Gould, Cleveland, Ohio). Under direct fluoroscopy, the electrode was directed to the vicinity of the bundle of His as detected on the electrocardiogram (ECG) (Fig. 1, left) and a 50-watt discharge was delivered for two minutes to produce AV nodal block. After heart block was confirmed (Fig. 1, right), the left jugular vein was exposed and a unipolar pacemaker lead was placed in the right ventricle under fluoroscopic visualization. A pulse generator (Medtronic Spectrax or Minix, Minneapolis, Minnesota), modified to pace at high heart rates, was programmed to beat asynchronously at 100 bpm and implanted in the subcutaneous tissue over the back of the neck.
Three additional dogs could not be successfully ablated and were treated as time controls. No differences in diastolic transmitral, pulmonary venous, or LA appendage flow velocities between baseline and eight weeks were noted (data not shown).
Right atrial (RA) pacemaker implantation
Three to five days after the RF ablation, dogs were anesthetized with morphine sulfate (3 mg/kg subcutaneously) and pentobarbital (20 mg/kg IV), and a unipolar pacemaker lead was sutured to the RA appendage through a right lateral third interspace thoracotomy and small pericardiotomy. The pulse generator was implanted into the subcutaneous tissue over the back of the neck.
Serial echocardiographic Doppler and hemodynamic studies
Five to seven days after RA pacemaker implantation, animals were lightly anesthetized as described above. The heart rate was slowed with 10 to 20 mg of DKAH 0264 (Boehringer Ingelheim, Ridgefield, Connecticut), which is an agent that inhibits the SA nodal If current without effects on myocardial contractility (14). A triple-lumen thermodilution catheter was advanced via a right jugular vein cut down to the pulmonary artery for measurement of cardiac output and RA and pulmonary artery pressure. Both atrial and ventricular pacing batteries were surgically exposed, and the pacing leads were connected to an AV sequential pacer set at 80 beats/min with an AV interval of 200 ms.
A biplane transesophageal imaging transducer (Hewlett-Packard, Palo Alto, California) was introduced into the esophagus and advanced to the level of the left atrium. The left atrial appendage (LAA), left upper pulmonary vein (PV), and mitral valve (MV) were visualized in transverse and orthogonal longitudinal “apical” views. Color-flow directed Doppler was used to identify intracavitary flow in all instances. The LAA flow was interrogated by positioning the sample volume in the area of maximal anterograde appendage flow. The sample volume was placed within the LV between the opened mitral leaflet tips, and 1 to 2 cm proximal to the LA in the left PV for interrogation of transmitral and pulmonary venous flow, respectively. Attempts were made to maintain the angle between the ultrasound beam and various flows within 30° using either apical view, and care was taken to make measurements from similar views during the serial studies. Four- and two-chamber views of the heart were recorded for measurements of LA and LV dimensions and mitral regurgitant jet areas.
Fluid-filled catheters were connected to Statham 23dB pressure transducers with zero pressure set at the level of the mid-right atrium. Temperature was monitored by the thermistor at the tip of the pulmonary artery catheter, and a table warmer was used to maintain body temperature at 38°C. Supplemental oxygen was administered to maintain a normal arterial blood PO2.
After the study was completed, the catheter was removed, the vascular insertion site repaired, and the pacemaker electrodes reconnected to the appropriate (i.e., atrial or ventricular) battery.
Doppler echocardiographic studies were digitized off-line with a dedicated image analysis system (Tomtec, Munich, Germany) and custom software. Diastolic transmitral waveforms were analyzed for the peak and integrated early (E) and late (A) velocities, and deceleration time. The peak E/A ratio, and the ratio of the A wave to total diastolic velocity time integral (Avti/diastolicvti) were calculated. Mitral deceleration was measured as the time between the peak early velocity and the point where the linearly extrapolated deceleration slope crossed the baseline. The LAA flow waveforms were analyzed for peak early (e) and late emptying (a) velocities, and their ratio (e/a) was derived (15). Pulmonary venous waveforms were analyzed for the PV diastolic reversal (during atrial systole), and peak and integrated systolic (J) and diastolic velocities (K). Integrals were calculated by dropping a vertical line to the baseline from the intersection of the systolic and diastolic waveforms. The fractional pulmonary venous velocity time integral (VTI) during ventricular systole (Jvti/[Jvti+Kvti]) was taken as the reservoir function, and the fractional pulmonary venous VTI during diastole (Kvti/[Jvti+Kvti]) was taken as the conduit function (13). The largest variance-encoded area of the mitral regurgitant jet was used to estimate the severity of mitral regurgitation (16).
Two-dimensional LV echocardiograms were planimetered at end-diastole (LVEDD) and end-systole as the maximum and minimum cavity dimensions, respectively. The LV ejection fraction was calculated using a single-plane Simpson’s rule algorithm. The LAed area was taken as the largest area immediately preceding the electrocardiographic P wave; because the entire LA could not be visualized owing to its location in the near field, a portion of the upper one-third of the LA was planimetered by extrapolation. In addition, the maximum LA dimension from either four- or two-chamber views was measured (13).
Thermodilution cardiac output determinations were made in (at least) triplicate and averaged. The pulmonary artery (PA) and RA pressures were electronically meaned. Studies were recorded at slow and rapid paper speeds (10 and 100 mm/s) on a Grass 7D multichannel recorder (Grass Instruments, Quincy, Massachusetts). Echocardiographic data were recorded on 0.5-in (1.27-cm) VHS videotape.
Data are expressed as mean ± SD. The effects of atrial and LV dysfunction on hemodynamic and echocardiographic measurements at a paced rate of 80 beats/min were compared using one-way repeated measures ANOVA (SigmaStat, SPSS). Tukey’s post hoc tests were used to determine where differences among significant effects were located. A p value < 0.05 was considered statistically significant.
Hemodynamic data at matched, AV sequentially paced heart rates are presented in Table 1. Mean RA and PA pressures increased significantly, and CO decreased significantly only after combined atrial and ventricular dysfunction.
Two-dimensional echocardiographic data
Echocardiographic data are presented in Table 2. The LV end-diastolic area increased and LV ejection fraction decreased significantly with isolated LV dysfunction, but was unchanged with isolated atrial systolic failure. The LV end-diastolic area increased and LV ejection fraction decreased with simultaneous atrial and ventricular dysfunction compared to baseline and isolated atrial failure, but was similar to values observed with isolated LV dysfunction. Left atrial end-diastolic area increased significantly only after combined LA and LV dysfunction; the similar trend observed in the LA end-diastolic dimension failed to achieve statistical significance (p = 0.09). The small increases in mitral regurgitant jet areas after isolated LV dysfunction and combined LA and LV dysfunction were small and not statistically significant.
Mitral Doppler data
Doppler echocardiographic data for the group are presented in Table 3, and examples from one of the animals are illustrated in Figures 2 through 4. The transmitral E/A ratio (Fig. 2) increased significantly after atrial pacing, consistent with a loss of atrial systolic function. The E/A ratio was less after LV dysfunction, but the difference was not statistically significant. To evaluate more critically the importance of atrial systolic function, the atrial systolic contribution (Avti) was expressed as a percentage of total diastolic VTI (Avti/diastolicvti). Isolated LA and combined dysfunction, and isolated LV dysfunction produced highly significant (and directionally opposite) changes in the percentage of atrial systolic flow. Thus, the absence of an effective (and compensatory) atrial systole was evident in animals with combined atrial and ventricular dysfunction.
Diastolic deceleration time was decreased after both isolated and combined atrial dysfunction, suggesting a decrease in net AV compliance. However, there was only a tendency (p = 0.08) for deceleration time to decrease after LV dysfunction was produced.
Left atrial appendage Doppler data
The LAA e/a ratios (Fig. 3) were significantly increased after both isolated and combined atrial dysfunction, but the decrease in e/a ratio after LV dysfunction failed to achieve statistical significance (p = 0.08).
Pulmonary flow Doppler data
The effects of atrial and ventricular dysfunction on pulmonary venous flow Doppler are shown in Table 3 and Figure 4. The PV diastolic reversal velocity during atrial systole was absent in both isolated and combined atrial failure, and significantly increased after isolated LV dysfunction, confirming the findings on mitral and LAA Doppler. Compared to baseline, the J/K ratio increased significantly with isolated LV dysfunction, and was unchanged in both isolated and combined atrial failure. However, when expressed as a function of the total PV flow, the conduit function [Kvti/(Jvti+Kvti)] increased significantly with isolated and combined atrial failure. Thus, in the absence of a compensatory atrial systole, the redistribution of reservoir and conduit functions was insufficient to maintain cardiac output. Indeed, total PV VTI was significantly decreased only in the animals with combined dysfunction.
Atrioventricular nodal ablation and rapid atrial and ventricular pacing were used to generate models of isolated atrial and ventricular dysfunction and their combination. The ability to pace the respective cardiac chambers independently, and to study reversible dysfunction in random and serial fashion, are unique aspects of this investigation. The principal findings of this study are 1) in dogs with normal LV systolic function, the loss of effective atrial systole does not alter cardiac output because of a compensatory increase in atrial conduit function; and 2) in dogs with chronic LV dysfunction, the redistribution of reservoir and conduit functions are insufficient compensation for the loss of effective atrial systole, and cardiac output falls. The study also confirms our earlier finding that isolated LV dysfunction produces compensatory changes in atrial booster pump function (17), and is similar to a report that found reduced LV ejection fraction without a change in cardiac output in dogs after three weeks of rapid (240 beats/min) ventricular pacing (18).
Atrial booster pump function
Pacing-induced atrial systolic failure was evident by the dramatic increase in the early to late diastolic transmitral and appendageal flow ratios. Although ejection phase indices of atrial systolic function were not measured in this study, previous work from this laboratory has shown that one week of rapid atrial pacing produces a marked impairment of left atrial systolic shortening without LV dysfunction (13). Moreover, very short-term rapid atrial pacing is known to produce reversible atrial systolic dysfunction (19).
Altering the duration and rate of ventricular pacing has been used to develop animal models with varying degrees of LV dysfunction (17,20–22). In the present study, a moderate degree of LV dysfunction was generated, characterized by an absence of overt heart failure, negligible increases in pulmonary artery pressure, and a moderately depressed LV ejection fraction. Nevertheless, the compensatory atrial changes were similar to those associated with a model of more severe LV dysfunction and congestive failure (17). In a serial study of microembolic heart failure, evolving heart failure was associated with an increase in the LA contribution to LV filling, whereas with progression, atrial systolic failure ensued (8). Thus, these data suggest that atrial systole may play a significant compensatory role during the development of LV dysfunction and heart failure.
Reservoir and conduit functions
Although the relative conduit function increased in both isolated and combined LA dysfunction, the absolute amount of PV flow (i.e., the totalvti), and hence the absolute conduit volume, decreased only with combined dysfunction. Thus, cardiac output falls because atrial systolic function is unable to compensate for LV dysfunction, and atrial conduit function is unable to compensate for atrial systolic failure.
Conduit function occurs primarily, but not exclusively, during ventricular diastole and represents the volume of blood that passes through the LA that cannot be attributed to reservoir or booster pump functions. In a well-characterized model, atrial conduit function accounted for 35 ± 8% of all flow through the atrium (3). Because minimum atrial stiffness is ∼four fold greater in the LA than ventricle, conduit function is favored during ventricular diastole (23). In the current study, net AV compliance was estimated by the transmitral deceleration time (24). The reduction in deceleration time (and hence compliance) was significant in response to rapid atrial pacing, and is consistent with the increase in atrial conduit function.
Several limitations of this study should be noted. First, “reversibility” of the models was determined by Doppler indices, which may have missed residual hemodynamic and mechanical effects. To minimize this bias, the sequence of LA and LV dysfunction was randomized. Second, there were no load-independent measures of LA function. Although dimensions and pressures are required for the accurate measurement of LA and LV diastolic function, Doppler mitral and LAA flows are useful surrogates for atrial contractile function, and net atrioventricular compliance can be estimated with the deceleration time (24), and the relative reservoir and conduit functions can be estimated with PV Doppler (1). Notably, the study exploits models with chamber-specific myocardial contractile dysfunction. Thus, the often difficult and important distinction between atrial myocardial failure and atrial systolic function owing to increased LA afterload is avoided. Third, caution should be urged when extrapolating from the lightly anesthetized dog to the conscious human; the models of atrial systolic failure and chronic LV dysfunction, and especially their combination, may not accurately reproduce features seen in human heart failure.
Moreover, filling patterns in the normal dog are most comparable to those seen in young, healthy human adults. The situation in the older patient, with a greater dependence on atrial contraction, may more closely resemble that observed in dogs with LV dysfunction—that is, the changes in conduit function may not be sufficient to compensate for an impairment in atrial systolic function, and decreased cardiac output may result. However, these pacing models permit a systematic analysis of isolated atrial and ventricular dysfunction and their interaction that would otherwise be impossible.
Recognizing the limitations inherent in this study, it seems reasonable to conclude that in the presence of a normal LV, impairment of atrial contraction has little effect on cardiac output, because conduit function compensates for atrial and ventricular filling. In contrast, in the presence of early LV dysfunction, when atrial booster pump and reservoir functions are increased (such as may occur with LV diastolic dysfunction), impairment of atrial contraction causes a decrease in cardiac output because atrial conduit function is unable to provide compensatory atrial and ventricular filling. These findings have implications for the need (based on cardiac mechanics) to convert atrial fibrillation to normal sinus rhythm and to implant AV sequential pacemakers.
We acknowledge the technical assistance of Harvey Hahn, MD, and Tom Friede. We also thank Boehringer Ingelheim for its generous gift of DKAH 0264.
☆ This work was supported in part by a Grant-in-Aid award from the American Heart Association (93006860) and with funds contributed in part by the AHA, Ohio–West Virginia Affiliate.
- left atrial (LA) reservoir function
- left atrial conduit function
- left atrium/left atrial
- left atrial appendage
- left ventricular (LV) end-diastolic dimension
- mitral valve
- pulmonary artery
- pulmonary vein
- right atrial
- velocity-time integral
- Received January 14, 2000.
- Revision received April 27, 2000.
- Accepted June 21, 2000.
- American College of Cardiology
- Hoit B.D,
- Shao Y,
- Tsai L.M,
- Patel R,
- Gabel M,
- Walsh R.A
- Hoit B.D,
- Shao Y,
- Gabel M,
- Walsh R.A
- Suga H
- Kono T,
- Sabbah H.N,
- Rosman H,
- Alam M,
- Stein P.D,
- Goldstein S
- Rahimtoola S.H,
- Ehsani A,
- Sinno M.Z,
- Loeb H.S,
- Rosen K.M,
- Gunnar R.M
- Matsuda Y,
- Toma Y,
- Ogawa H
- Miura T,
- Miyazaki S,
- Guth B.D,
- Kambayashi M,
- Ross J Jr.
- Hoit B.D,
- Shao Y,
- Gabel M
- Hoit B.D,
- Jones M,
- Eidbo E.E,
- Elias W,
- Sahn D.J
- Hoit B.D,
- Shao Y,
- Gabel M,
- Walsh R.A
- Komamura K,
- Shannon R,
- Ihara T,
- et al.
- Leistad E,
- Christensen G,
- Ilbekk A
- Spinale F.G,
- Holzgreefe H.H,
- Mukherjee R,
- et al.
- Spinale F,
- Holzgrefe H,
- Mukherjee R,
- et al.
- Alexander J,
- Sunagawa K,
- Chang N,
- Sagawa K
- Thomas J,
- Newell J,
- Choong C.Y,
- Weyman A