Effects of Exercise During Long-term Support With a Left Ventricular Assist Device
To the Editor:
Jaski et al,1 in their article in Circulation, provide important information regarding exercise physiology during long-term TCI Heartmate left ventricular assist device (LVAD) support. They indicate that the native left ventricle contributes directly to systemic cardiac output during exercise yet do not attempt to quantify left ventricular contribution.
In our institution, of more than 660 patients referred for cardiac transplantation between 1990 and 1997, 50 patients had to receive LVAD support, 39 by the Novacor,2 7 by the TCI Heartmate, and 4 by the Medos system. We have reported a protocol for the Novacor LVAD that allows for quantitation of native left ventricular function.3 A similar protocol was developed for the TCI Heartmate LVAD with the purpose of (a) quantifying left ventricular contribution and (b) assessing left ventricular functional reserve. A 7.5F thermodilution Swan-Ganz catheter was placed in the pulmonary artery. Cardiac output, pulmonary capillary wedge pressure, pulmonary artery pressure, and right atrial pressure as well as arterial pressure (by cuff inflation) were recorded at rest, during upright bicycle exercise after 5 minutes on the 25-, 50-, and 75-W levels as well as 3 and 10 minutes after exercise. Simultaneously, TCI Heartmate LVAD hemodynamic data (pump rate, pump stroke volume, pump output) were automatically recorded. Data were compared by subtraction. Two different LVAD settings were chosen. In setting 1 (“fill-to-empty”), we achieved maximal pump washout in conjunction with optimal pump output, as did Jaski et al.1 In this setting, the aortic valve remained largely closed as assessed by simultaneous transthoracic echocardiography. The LVAD acted as a series pump. The left ventricle was unloaded. Setting 2 (“fixed rate mode”) consisted of a fixed pump rate of 50/min so that synchronization of pump and native left ventricular action was no longer possible. In this situation, incomplete emptying of the left ventricle into the pump was encountered with a resultant rise in left ventricular load. Thus the pump rate reduction actually led to an afterload and preload challenge for the native left ventricle. In this situation, total cardiac output even at rest was substantially supported by the recipient’s own heart as documented by aortic valve opening. The LVAD acted as a parallel pump.
This protocol was, as a pilot test with institutional review board approval and patient informed consent, applied in the last 3 of our 7 TCI Heartmate recipients (all men; mean age, 50±12 years; 1 idiopathic, 2 ischemic cardiomyopathy) after recovery (117±20 days after surgery) and informed consent. At the time of the study, all patients were receiving vasodilators, aspirin, and dipyridamole. Results are summarized in the Table⇓ . The main finding was the quantification of the percentage left ventricular stroke volume across the aortic valve (between zero and 53% of total=right ventricular stroke volume) and a higher percent left ventricular stroke volume (53±14% versus 15±12%; P=.024) on the 75-W level in setting 2 compared with setting 1. The difference may be considered as left ventricular functional reserve.
It has been shown that exercise testing in LVAD recipients can be safely performed.4 Patients with the TCI left ventricular assist device walked up to 6 miles and were in New York Heart Association class II.5 Functional benefit with increase of peak O2 to 15 mL/kg per minute and normalization of neurohormones has been reported.6 In the Pittsburgh series, in 31 Novacor patients, mean peak O2 increased from 10 to 15 mL/kg per minute.7 All of these LVAD exercise protocols have assessed the combined function of the left heart–left ventricular assist device complex without quantifying the specific contribution of the native left ventricle.
With the proposed protocol, the relative contribution of the native left ventricle to total stroke volume and cardiac output may be quantified. Furthermore, left ventricular functional reserve can be safely and reliably assessed in clinically stable TCI Heartmate patients. Thus this protocol may facilitate prediction of left ventricular recovery and the potential for weaning, the potential for chronic LVAD support, as well as estimation of the risk associated with device dysfunction.
- Copyright © 1998 by American Heart Association
Jaski BE, Kim J, Maly RS, Branch KR, Adamson R, Favrot LK, Smith SC, Dembitsky WP. Effects of exercise during long-term support with a left ventricular assist device. Circulation.. 1997;95:2401–2406.
McCarthy PM, James KB, Savage RM, Vargo R, Kendall K, Harasaki H, Hobbs RE, Pashkow FJ, and the Implantable LVAD Study Group. Implantable left ventricular assist device: approaching an alternative for end-stage heart failure. Circulation. 1994;90(suppl 2):II-83–II-86.
We appreciate Dr Deng and his coworkers for acknowledging the findings of the EVADE pilot trial and for providing additional data in their carefully studied three patients. We also found that from rest to peak exercise, Fick systemic blood flow increased significantly more than left ventricular assist device (LVAD) output (2.8±1.9 versus 1.6±1.1, P<.05), implying parallel ejection of blood through the native aortic valve in the fill-to-empty mode used for our study. The fixed-rate protocol used by Dr Deng and coworkers for quantifying LV contribution may serve as a valuable tool either for predicting LV recovery before LVAD explant or for evaluating patient safety during chronic LVAD support. Both of our studies address the need for additional data from larger prospective studies to help understand the unique physiology of the LV-LVAD complex as well as to further define the role of the LVAD in long-term implantation.
We disagree with the calculation of percent LVSV through an LVSV/RVSV ratio, because it implies that LVAD filling occurs only during LV systole. Because LVAD filling also occurs during LV diastole in response to the negative pressure induced by the LVAD during its filling phase,R1 the actual percent LVSV ejected through the native aortic valve during LV systole should be, in fact, higher than the values reported in their study.
Branch KR, Dembitsky WP, Peterson KL, Adamson R, Gordon JB, Smith SC Jr, Jaski BE. Physiology of the native heart and Thermo Cardiosystems left ventricular assist device complex at rest and during exercise: implications for chronic support. J Heart Lung Transplant.. 1994;13:641–651.