| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J Thorac Cardiovasc Surg 2004;128:98-102
© 2004 The American Association for Thoracic Surgery
Cardiopulmonary support and physiology |
a Department of Surgery, Columbia University College of Physicians and Surgeons, New York, NY, USA
b Department of Biostatistics, Columbia University College of Physicians and Surgeons, New York, NY, USA
Received for publication August 5, 2003; revisions received December 10, 2003; accepted for publication January 21, 2004.
* Address for reprints: Henry M. Spotnitz, MD, Department of Surgery, Columbia University College of Physicians and Surgeons, 622 W 168th St, PH 14-103, New York, NY 10032, USA
hms2{at}Columbia.edu
 |
Abstract |
|---|
 Top Abstract Materials and methodsResultsDiscussionReferences |
|---|
METHODS: After median sternotomy in 6 anesthetized domestic pigs, complete heart block was induced by ethanol ablation. A conduit was grafted from the right ventricle to the right atrium to simulate tricuspid insufficiency. During epicardial, atrial tracking DDD biventricular pacing, atrioventricular delay was varied between 60 and 180 ms in 30-ms increments. Right ventricular–left ventricular delay was varied at each atrioventricular delay from +80 ms (right ventricle first) to –80 ms (left ventricle first) in 20-ms increments. Aortic flow, right ventricular pressure, and electrocardiogram were measured at each pacemaker setting with the graft clamped and unclamped.
RESULTS: Atrioventricular and right ventricular–left ventricular delays had no significant effect on cardiac output with the graft clamped. With the graft unclamped, however, there was a statistically significant (P = .003 by mixed modeling repeated measures analysis of variance) trend toward higher cardiac output with left ventricle–first pacing.
CONCLUSION: Left ventricle–first biventricular pacing in swine significantly increased cardiac output during acute tricuspid insufficiency but not during the control state. Trials are warranted to develop clinical biventricular pacing for treatment of perioperative right ventricular dysfunction.
A previous study from our laboratory of BiVP in swine examined effects of atrioventricular delay (AVD) and interventricular pacing delay (right-left delay, RLD) on cardiac output (CO) in critical pulmonary stenosis with RV pressure overload.1,2 CO was optimized by a positive RLD, the RV paced before the LV. This experiment has extended that study to a different type of heart failure, RV volume overload, achieved by simulating tricuspid insufficiency (TI) with an RV–right atrial (RA) conduit in an open-chest porcine model.
Clinical studies demonstrate that BiVP can enhance stroke volume in patients with congestive heart failure3 and that the effect is RLD dependent. However, the optimum RLD is highly variable from patient to patient and from time to time in the same patient.4-6 We speculate that this variability reflects variation in right-sided and left-sided pressure or volume overload. Our studies were designed to examine this possibility by evaluating the effect of RLD on stroke volume during isolated forms of heart failure.
|
Materials and methods |
|---|
|
TopAbstractMaterials and methodsResultsDiscussionReferences |
|---|
Median sternotomy was carried out in 8 anesthetized domestic pigs (35-45 kg). BiVP with variable RLD was tested with a bipolar RA sensing lead (Medtronic Inc, Houston, Tex) split into two current paths, each connected to a temporary external pacemaker (Medtronic 5388). Bipolar epicardial pacing leads were also placed on the anterior surface of the RV and posterior surface of the LV. Specifically, the RV lead was placed in the center of the anterior RV toward the RV outflow tract in a vessel-free region. The LV lead was placed by mobilizing the heart anteriorly and suturing the lead to the posterior surface of the LV midway between apex and base. The pacemakers were programmed to minimum ventricular sensitivity (ventricular asynchronous) to prevent inhibition of the second pacemaker by the initial ventricular pacing artifact during BiVP with an RLD greater than 0.
Once proper function of the pacemaker leads was confirmed, complete heart block (CHB) was established by injection of 0.5-mL aliquots of 100% ethanol into the region of the bundle of His at the base of the aorta. The cumulative amount of ethanol required to establish CHB ranged from 0.5 mL to 5 mL. In 2 of the 8 animals, during the attempt to establish third-degree heart block, the RV promptly dilated, systemic blood pressure plummeted, and the hearts showed a fine ventricular fibrillation. One animal could not be cardioverted, and the other appeared globally hypokinetic after electrical cardioversion. These conditions are presumed to reflect intracoronary ethanol injection, and both animals were excluded from data analysis. After CHB was established, a segment of thoracic aorta harvested from another animal and stored in Ringer's lactate and 2% formalin solution was rinsed and then used as a graft. Modified right-angle aortic arch cannulas (6.5 mm Sarns; 3M Health Care, Ann Arbor, Mich) were secured to each end of the graft. The metal cannulas were introduced sequentially through 4-0 Prolene purse-string sutures (Ethicon, Inc, Somerville, NJ) first into the RV apex and then, after confirmation of vigorous flow, into the RA. In 1 animal an ultrasound transit-time flow probe (Transonic Systems Inc, Ithaca, NY) placed around the graft confirmed flow, which measured 15% to 20% of CO.
The pacing protocol was atrial tracking DDD BiVP. The AVD was varied from 60 to 180 ms in 30-ms increments. At each AVD the RLD was varied from +80 ms (RV first) to –80 ms (LV first) in 20-ms increments as described previously elsewhere.1 At each pacing setting, ECG, aortic flow, RV pressure, and peripheral arterial pressure were measured in the control state with the graft clamped and during TI with the graft open. This order was not varied between experiments. After data were collected, we again clamped the graft and measured CO to assess the stability of the model. Animals were humanely put to death at the conclusion of the experiment.
Data analysis
ECG, RV pressure, and aortic flow velocity were sampled at 200 Hz with a 16-channel analog to digital converter (MacLab; ADInstruments Inc, Milford, Mass) and recorded on a digital computer (Power Macintosh 7100/66; Apple Computer, Inc, Cupertino, Calif). CO was determined for each experimental phase by integrating aortic flow velocity over time during a single complete respiratory cycle free of arrhythmia. A statistical modeling procedure (described in the following section) was used to determine the effects of heart rate, AVD, and RLD on CO. To determine the effect of AVD on CO, the RLD was kept at 0 and the AVD was changed while the aortic flow velocity was recorded. Heart rate was calculated from the average R-R interval of 5 beats free of ectopy.
Statistical analysis
Changes in CO across various levels of AVD were modeled with the PROC MIXED procedure in the SAS software suite (SAS Institute, Inc, Cary, NC). Because repeated measurements within an animal may be correlated, this procedure allows one to model this correlation structure, commonly referred to as a covariance pattern. This allows improved estimates of the standard errors of measurement and thus more powerful tests. There are a number of covariance structures available. Three of the more common ones are compound symmetry, for correlations that are constant for any two points in time, autoregressive order one, for correlations that are smaller for time points further apart, and unstructured, which has no mathematic pattern within the covariance matrix. Other covariance structures that are usually tested include the Toplitz and the heterogeneous compound symmetry structure. A likelihood ratio test or a procedure known as the Akaike information criterion is used to discern which covariance pattern allows the best fit.7
|
Results |
|---|
|
TopAbstractMaterials and methodsResultsDiscussionReferences |
|---|
|
|
|
|
|
Discussion |
|---|
|
TopAbstractMaterials and methodsResultsDiscussionReferences |
|---|
This study did not examine the importance of lead location relative to RLD. That will be the subject of a future study. In addition, the use of a valved conduit to simulate TI may have produced regional dysfunction of the anterior RV that is not present clinically.
Our previous study of pulmonary stenosis demonstrated that CO increased 20% during optimized pacing versus simultaneous BiVP.1 Optimum RLD in that model of RV pressure overload was +40 ms, the RV paced before the LV. In this experiment, LV-first BiVP provided the maximum CO, but the improvement versus simultaneous BiVP was smaller than in our previous study and the timing was less reproducible.
Thus 1 of our 6 animals in TI responded differently than the others despite normal baseline hemodynamics. This variability may reflect problems with our model, including possible ethanol-related myocardial injury, injury to the RV during grafting of the RV-RA conduit, or inconsistent regurgitant flow through the conduit despite our attempts to standardize the degree of TI. There was consistent depression of CO after establishment of CHB and insertion of the RV-RA conduit, suggesting a possible generalized injury. Our conclusions concerning the effect of RLD on CO during TI remain valid, however, because whatever problems were present affected the control state as well. Despite less compelling results, this experiment advances our previous work to a new form of acute heart failure and increases the evidence that optimized timing during BiVP reflects pathologic loading conditions.
Interestingly, although there was no statistically significant effect of AVD on CO during TI, there was a trend towards improved output at higher AVDs. It appears likely that reduction of CO in this study reflects RV failure and that effects of AVD and RLD are mediated by changes in RV function. Meisner and colleagues8 defined optimal AVD as the interval allowing the atrioventricular valves to remain open long enough to maximize ventricular filling without permitting atrioventricular regurgitation.
Our publications to date have not defined mechanisms for the effects of AVD and RLD on CO and have not examined mechanistic differences between effects in pulmonary stenosis or TI. AVD is believed to act through effects on ventricular filling, whereas RLD is more likely to influence critical aspects of the synchrony of contraction. Timing is likely to have important effects on end-diastolic volume in TI because of the critical balance between antegrade and retrograde flow. On the other hand, in critical pulmonary stenosis it is likely that the RV is maximally distended by afterload mismatch, and RV-first pacing may provide advantages by increasing the duration of effective systolic ejection. Measurements of atrial filling pressures as well as RV and LV pressures, volume, shape, and wall motion are critical to understanding these issues. Our preliminary data indicate that optimized RLD improves LV symmetry throughout the cardiac cycle.2
Studies of the mechanism of optimization of BiVP are clearly needed. Such studies should define geometric changes by means of echocardiography, sonomicrometry, or conductance. These methods have been used to advantage by our laboratory in studies of pulmonary stenosis.2,9 Myocardial oxygen consumption would also be an interesting subject of future research.
Our results provide insight into potential avenues for clinical development of BiVP. Our previous studies in pulmonary stenosis indicate not only that RV first pacing is beneficial, but also that properly timed BiVP is better than either RV or LV pacing alone. These results in TI also indicate that BiVP is preferable to LV or RV single-site pacing, with LV pacing preceding RV pacing for optimum benefit. Although acute studies in laboratory animals with normal myocardium may not be directly transferable to the clinical environment, accumulating data suggest that patients in acute heart failure are likely to benefit from BiVP rather than single-site LV or RV pacing.10 The best results are likely to be obtained with timing optimized for each patient. Clinical studies are needed to confirm this hypothesis.
|
Footnotes |
|---|
*George H. Humphrey, II, Professor of Surgery.
|
References |
|---|
|
TopAbstractMaterials and methodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  T. A. Quinn, G. Berberian, S. E. Cabreriza, L. J. Maskin, A. D. Weinberg, J. W. Holmes, and H. M. Spotnitz Effects of sequential biventricular pacing during acute right ventricular pressure overload Am J Physiol Heart Circ Physiol, November 1, 2006; 291(5): H2380 - H2387. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Load dependence of cardiac output in biventricular pacing: left ventricular volume overload in pigs. J. Thorac. Cardiovasc. Surg., March 1, 2006; 131(3): 666 - 670.
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
