Potential of left ventricular assist devices as outpatient therapy while awaiting transplantation

Potential of Left Ventricular Assist Devices as Outpatient Therapy While Awaiting Transplantation Howard R. Levin, MD, Jonathan M. Chen, MD, Mehmet C. OZ, MD, Katharine A. Catanese, RN, Henry Krum, MD, PhD, Rochelle L. Goldsmith, PhD, Milton Packer, MD, and Eric A. Rose, MD Division of Circulatory Physiology, Department of Medicine, and Division of Cardiothoracic Surgery, Department of Surgery, Columbia University College of Physicians and Surgeons, New York, New York

Left ventricular assist devices (LVADs) increasingly are being used as a bridge to transplantation. We studied changes in New York Heart Association class, mean arterial pressure, resting cardiac output, end-organ function, exercise oxygen consumption, and exercise cardiac output in 12 LVAD recipients. In addition, resting levels of neurohormonal factors were evaluated 4 to 16 weeks after implantation. Two of the 12 patients died of right heart failure and 1 of aspiration; all deaths occurred in the first 2 weeks after LVAD implantation. Of the other 9 patients, 8 improved to New York Heart Association class I and 1 to class 11, all of whom were in class IV preoperatively. The 4 patients who underwent exercise testing achieved an exercise oxygen consumption of 15.0 ± 2.7 mL· kg- 1 • min- 1 , which was paralleled by an increase in resting cardiac output from 3.07 ± 0.9 L'min- 1 preoperatively to 5.66 ± 1.1 L' min"? at 2

months, and mean arterial pressure from 60 ± 8 to 91 ± 10 mm Hg at 2 months, a benefit that was maintained for up to 10 months. End-organ function revealed comparable improvement at 2 months for both creatinine (1.68 ± 0.7 to 1.0 ± 0.19 mg . dL -1) and total bilirubin (1.37 ± 1.17 to 0.54 ± 0.26 mg . dL -1) levels. Levels of neurohormones were within normal limits. Adverse clinical events after the perioperative period were minimal, and no thromboembolic complications occurred. These data indicate that functional and physiologic recovery occurs during LVAD support, that adverse clinical and mechanical events occurring after the perioperative period are few and should not preclude hospital discharge in selected patients, and thus that LVADs can provide reliable, longterm support and should be evaluated for future use as outpatient therapy. (Ann Thome Surg 1994;58:1515-20)

W

dynamics, and exercise capacity in some patients with severe heart failure [4]. Thus, there has been an increased use of LVADs to support the failing circulation of patients awaiting transplantation. Left ventricular assist devices have been used successfully as a "bridge to transplantation" in the inpatient setting for up to 345 days; moreover, hospital stays in excess of 3 months frequently have been described [5, 6]. Without question, such extended continuous inpatient hospitalizations, with their attendant high costs, impaired quality of life, and increased risks of infection, represent a burden to both the patient and the health care system. Previous investigators have reported an improved quality of life in a small group of selected "bridged" recipients who were allowed to await transplantation at a designated outpatient center [7]. However, it is unknown whether sufficient functional recovery would develop in all LVAD recipients to allow them to be sent home safely while awaiting transplantation. Therefore, we undertook the present investigation both to characterize the clinical and physiologic response to long-term LVAD support and to identify potential factors that would preclude successful temporary discharge while awaiting transplantation.

ith a prevalence of more than 2,300,000 nationwide, and more than 400,000 new cases recorded each year, chronic congestive heart failure (CHF) represents a serious condition for which improved treatment modalities are needed urgently [1]. Although cardiac transplantation offers an effective alternative for many patients with end-stage CHF, the lack of donors currently limits the number of For editorial comment, see page 1309. transplantations performed annually nationwide to approximately 2,000, or only 5% of the potential recipients [2]. As a consequence, 20% of patients awaiting transplantation die before they can receive a suitable donor heart [3]. Current clinical data suggest that left ventricular assist devices (LVADs) can improve end-organ function, resting hernoAccepted for publication May 17, 1994. Presented at the Sixty-sixth Scientific Sessions of the American Heart Association, Atlanta, GA, November 9,1993. Address reprint requests to Dr Levin, Division of Circulatory Physiology, Columbia-Presbyterian Medical Center, 177 Fort Washington Ave, MHB 5-435, New York, NY 10032.

© 1994 by The Society of Thoracic Surgeons

0003-4975/94/$7.00

1516

LEVIN ET AL LEFf VENTRICULAR DEVICE AS OUTPATIENT THERAPY

Connection to External Air Power Supply

Fig 1. Schematic representation of the left ventricular assist device used in the present study. (Reprinted from Cardiac Chronicle 1993;7: 1-7, by permission.)

Material and Methods

Description of the Device The TCI Heartmate 1000 IP LVAD used in the present study is of pusher-plate design, with a maximum stroke volume of 85 mL. The device is implanted through a median sternotomy incision (Fig 1) using apical cannulation to maximize left ventricular unloading.

Operative Implantation Before heparinization, a properitoneal plane is developed, and after systemic heparinization (3 mg' kg-I), the aorta and right atrium are cannulated and cardiopulmonary bypass is instituted. A left ventricular vent is placed through the right superior pulmonary vein and, without placement of a cross-clamp or administration of cardioplegia, device implantation is initiated. The driveline is tunneled to an exit point and the inflow cannula is brought through the diaphragm and secured to a Teflon cuff attached to the ventricular apex. The outflow graft is anastomosed to the right lateral aspect of the ascending aorta, and, as the patient is separated from cardiopulmonary bypass, the left ventricular vent is removed and the device is activated. After reversal of heparin with protamine, and with meticulous attention paid to hemostasis in these often coagulopathic patients, large chest tubes are placed within the mediastinum and properitoneal space, and the chest and linea alba are closed.

Patient Selection Patients selected for device implantation had to be approved transplant candidates who met hemodynamic criteria for LVAD implantation within 24 hours of implantation. Patients could be entered in the study if they had a pulmonary capillary wedge pressure greater than or equal

Ann Thorac Surg 1994;58:1515-20

to 20 mm Hg and either a systolic blood pressure less than or equal to 80 mm Hg or a cardiac index less than or equal to 2.0 L· min-I. m- 2 . Exclusion criteria were similar to those for transplantation. All investigations were undertaken as directed by protocols approved by the investigational review board at our institution. The study population consisted of 12 consecutive patients who had undergone LVAD implantation from August 1990 to May 1993. This cohort was predominantly male (92%), ranging in age from 18 to 66 years (mean, 49.3 years). Six patients had an idiopathic cardiomyopathy, whereas 6 patients had ischemic disease. Severe preoperative hemodynamic compromise was present as evidenced by a mean pulmonary capillary wedge pressure of 29.1 ± 9.8 mm Hg and a mean thermodilution cardiac output of 3.07 ± 0.93 L· min- I despite maximal inotropic support. These patients were followed up for 0 to 324 days of LVAD support (mean, 102 days). Two of the 12 patients died of right heart failure and 1 died of aspiration. All deaths occurred within 2 weeks of LVAD implantation. The remaining 9 patients represent the cohort evaluated as described in the protocol below. The basic study design is represented by Figure 2. The first 4 patients in the study received aspirin therapy; the subsequent 8 patients received no anticoagulant therapy.

Protocols Patients underwent classification by the New York Heart Association (NYHA) functional classification system both before and after LVAD implantation. Maximal treadmill exercise testing with gas exchange was administered 4 to 16 weeks (mean, 10 weeks) after LVAD implantation using a modified Naughton protocol. Breath by breath expired gas analysis was performed on-line during the test with a commercially available metabolic cart (Sensormedics, Yorba Linda, CA). The test was performed without interruption, and the patient was encouraged to exercise to the point of exhaustion. Peak oxygen consumption was defined as the highest oxygen consumption achieved during the test averaged over 20 seconds. FUNCTIONAL EVALUATION.

LVAD Patient

Adverse Events

/~

Functional

Physiologic

Clinical

Mechanical

NYHACIa•• Exercl.e C.peclty

Hemodynamic

RHF

Renal Hepatic

DeviceMalfunction

Thromboembolic

Neurohormonal

Arrhythmias Infection Hemolysis Bleeding

Fig 2. Schematic representation of the overall study design. Aspects of functional and physiologic outcome were evaluated in conjunction with potential adverse events that would preclude safe hospital discharge. (LVAD = left ventricular assist device; NYHA = New York Heart Association; RHF = right heart failure.)

PHYSIOLOGIC EVALUATION. Mean arterial pressure and resting cardiac output (as represented by LVAD outflow) were evaluated daily throughout the duration of mechanical assistance. Renal and hepatic function were evaluated at weekly intervals throughout the duration of LVAD support by serum creatinine, blood urea nitrogen, serum glutamic-oxaloacetic transaminase, serum glutamic-pyruvic transaminase, and total bilirubin concentrations. Levels of neurohormonal factors were assessed from 4 to 16 weeks after LVAD implantation in 4 patients. Patients had a 21-gauge intravenous cannula inserted into a forearm vein, then were allowed to rest in a supine position for 30 minutes. Twenty milliliters of blood was obtained and analyzed for neurohormones. Neurohormonal factors assessed included plasma norepinephrine as measured by high-pressure liquid chromatography), serum aldosterone (as measured by radioimmunoassay), and plasma renin activity (as measured by radioimmunoassay) (Smith Kline Beecham Clinical Laboratories, Syosset, NY). To compare these post-LvAD patients with their neurohormonal status before implantation, levels of neurohormonal factors were also measured in 8 non-LvAD NYHA class IV patients.

EVALUATION OF

ADVERSE

CLINICAL

AND

MECHANICAL

Adverse clinical and mechanical events that occurred during the period of LVAD support were evaluated retrospectively on the basis of the following criteria: (1) right heart failure was defined as that condition demonstrating the clinical need for mechanical right ventricular assistance, (2) arrhythmias were defined as sustained ventricular tachycardia or ventricular fibrillation, (3) thromboembolic episodes were defined as computed tomography-proven evidence of a thromboembolic event associated with a neurologic defect or transient event in the setting of appropriate clinical findings, (4) infections were defined as those blood or wound culture positive infections that required therapy, (5) persistent hemolysis was defined as a plasma free hemoglobin level greater than 5.0 mg' dL -1, (6) bleeding was defined as an event significant enough to require return to the operating room for further management, (7) renal insufficiency was defined as a serum creatinine level greater than 2.2 mg' dL -1 or a blood urea nitrogen level greater than 50 mg' dL -1, (8) hepatic failure was defined as serum total bilirubin greater than 1.4 mg' dL -1 or serum glutamic-oxaloacetic transaminase level greater than 50 U' L -1 or serum glutamicpyruvic transaminase level greater than 50 U' L -1, (9) mechanical events were defined as cessation of device function, and (10) other complications included decreased appetite and small bowel obstruction. Thirty days was arbitrarily chosen as the defining point for analysis because it represents the earliest time after which it was thought that patients potentially could be discharged home safely from the hospital. Thus, the incidence of adverse events was separated into two groups, less than or equal to 30 days (the perioperative period) or greater than 30 days after implantation. EVENTS.

1517

LEVIN ET AL LEFT VENTRICULAR DEVICE AS OUTPATIENT THERAPY

Ann Thorae Surg 1994;58:1515-20

"2 7

~

-~

/-~

6

~

5

o

4

C

~

...J

3

•

I-~--,--~-..,....--r--~-""""'--,--.---r-

120

Ci J: E

100

S

80

~

60

0.

•

I~I~ ----

-------I

1 30

90

150

210

324

Duration of LVAD Support (Days)

Fig 3. Hemodynamic response to left ventricular assist device (LVAD) support as represented by mean arterial pressure (MAP) and LVAD output evaluated throughout the time of LVAD support.

Results

Functional Evaluation Of the nine patients who survived the perioperative period, all were in NYHA class IV before LVAD implantation. Eight of 9 (89%) improved to NYHA class I and 1 of 9 (11%) improved to NYHA class II after LVAD implantation. Four patients underwent exercise testing for 10.2 to 19 minutes (mean, 12.3 ± 4.8 minutes). These patients achieved a peak oxygen consumption of 15.0 ± 2.7 mL· kg-I. min-I, attained 83.7 ± 13.2% of peak predicted heart rate and increased their cardiac output by 3.54 ± 0.97 L· min- 1 over preexercise values.

Physiologic Evaluation Resting cardiac output (as measured by pump output) was evaluated in all patients and was demonstrated to increase from 3.07 ± 0.9 L· min- 1 (range, 1.9 to 4.6 L· min-I) preoperatively (n = 12) to 5.66 ± 1.1 L· min- 1 (range, 4.4 to 7.5 L . min -1) at 2 months (n = 8) and 5.4 ± 1.3 L· min- 1 (range, 3.2 to 6.7 L· min-I) at 6 months (n = 4). Mean arterial pressure increased from 60 ± 8 mm Hg (range, 48 to 80 mm Hg) preoperatively (n = 12) to 91 ± 10 mm Hg (range, 84 to 107 mm Hg) at 2 months (n = 8) and 89 ± 11 mm Hg (range, 75 to 107 mm Hg) at 6 months (n = 4). These improvements in resting cardiac output and mean arterial pressure were maintained for up to 10 months (Fig 3).

Studies at weekly intervals throughout the duration of LVAD support revealed an improvement in serum creatinine values from 1.68 ± 0.7 mg' dL -1 (range, 1.1 to 3.5 mg' dL -1) preoperatively (n = 12) to 1.0 ± 0.2 mg' dL -1 at 2 months (n = 8) and 1.05 ± 0.05 mg' dL -1 (range, 1.1 to 1.6 mg' dL -1) at six months (n = 5) (Fig 4). Similarly, serum bilirubin total levels improved from 1.37 ± 1.17 mg' dL -1 (range, 0.4 to 4.9 mg' dL -1) preoperatively (n = 12) to 0.54 ± 0.26 mg . dL -1 (range, 0.3 to 1.1 mg' dL -1) at

1518

LEVIN ET AL LEFf VENTRICULAR DEVICE AS OUTPATIENT THERAPY

2.5

,...... ~ 2.0 Cl

§.

1.5

:.cj

1.0

c

:E iii

0.5

,...... ~

3

Cl

§.

0.0

2

Q)

C

·C ~

l!!

o

~_1~1_1_I_I_T I-~_._ •.. .- - - ._ _ _ ._.

0

30

90

150

210

•

• 324

Fig 4. End-organ response to left ventricular assist device (LVAD) support as represented by serum creatinine and total bilirubin levels evaluated throughout the time of LVAD support.

2 months (n = 8) and 0.46 ± 0.29 mg' dL -1 (range, 0.2 to 1.0 mg' dL -1) at 6 months (n = 5) (Fig 4). Plasma free hemoglobin levels were not recorded preoperatively, but at 2 months the mean level was 4.4 ± 4.0 mg' dL- 1 (normal, <5 mg' dL -1). In the 4 patients who underwent exercise testing, levels of aldosterone (7.8 ± 3.95 ng' dL -1) and plasma renin activity (1.9 ± 1.0 ng' ml- 1 • h- 1 ) were within normal limits (normal range: aldosterone, <16 »s dL -1, plasma renin activity, 0.2 to 2.3 ng' mL -1. h- 1 ) . Levels of norepinephrine (494 ± 194 pg' mL -1) were above the normal range (15 to 86 pg' mL -1) but were markedly lower than those measured in the class IV patients (831 ± 113 rs mL -1).

Evaluation of Adverse Clinical and Mechanical Events The adverse clinical events during LVAD support are summarized in Table 1. As shown, most adverse events occurred during the 30-day perioperative period, and only two episodes of arrhythmia (one ventricular fibrillation, one ventricular tachycardia), five treated infections (three Table 1. Frequency of Adverse Events" Adverse Event Right heart failure Arrhythmias Thromboembolic event Infection Hemolysis Bleeding Renal dysfunction Hepatic dysfunction Mechanical dysfunction Other

:0;30 Days

>30 Days

2 3

o

o 5

o 2

o 1 1 2

2

o 5

o o 2

o o 1

a Frequency of adverse events are listed for the 12-patient left ventricular assist device cohort. Events were subdivided arbitrarily into those occurring during the 3D-day perioperative period and those occurring after this time.

Ann Thorac Surg 1994;58:1515-20

of which involved line sepsis), two episodes of renal insufficiency, and one episode of decreased appetite occurred after the perioperative period. Only 1 patient had an infectious episode both during and after the 30-day perioperative period. Notably, there were no thromboembolic complications during either LVAD implantation or during subsequent circulatory support. All autopsies showed no thromboembolic events in the brain or other organs. One adverse mechanical event occurred during the perioperative period due to battery failure as a result of human error.

Comment Despite recent advances in the medical therapy, CHF still is the leading cause of morbidity and mortality in industrialized nations, with more than 4 million Americans reportedly afflicted in 1990, and an additional annual incidence of 400,000 patients [1, 8, 9]. In 1988 alone, CHF represented the principal cause of death in 37,000 patients and a contributing cause of death in 200,000 patients [10, 11]. Further, current data indicate a progressive increase in the number of hospital admissions for CHF throughout the last 20 years [12, 13]. Three major therapeutic modalities have emerged in the treatment of end-stage CHF: medical therapy, cardiac transplantation, and mechanical assist devices. Recent evidence suggests that the successful treatment of left ventricular systolic dysfunction or hypertension may significantly decrease the incidence of CHF [14-16]. However, the beneficial epidemiologic impact of these therapies has yet to be realized fully. Although cardiac transplantation has achieved approximately 85% one-year and 65% fiveyear survival rates nationwide, its success remains limited by the complications of chronic immunosuppression and a current lack of donor organs. As a result, there has been an increased interest in the potential therapeutic application of LVADs. Since they were first introduced by Cooley and associates [17] in 1969, temporary mechanical circulatory support devices have become useful as bridges-to-transplantation for those patients awaiting cardiac transplantation. Previous investigators have shown it to be possible for patients with LVADs to function well in an outpatient setting [7]. Therefore, we undertook the current investigation to evaluate the feasibility of temporary outpatient discharge while awaiting transplantation in all patients receiving LVAD support. For temporary discharge-- or permanent LVAD implantation-to be a practical therapeutic alternative, however, both functional ability and exercise capacity in recipients must be improved to a level that allows them to perform an acceptable level of activity. Left ventricular assist device support in our cohort significantly improved the recipients' NYHA functional class, a result consistent with previously documented results [4]. This level of overall exercise tolerance and functional ability suggests that LVAD support may benefit these patients' functional capacity to a degree sufficient to permit outpatient activities of daily living.

Ann Thorae Surg 1994;58:1515-20

The physiologic evaluation of our recipients revealed a consistent improvement in end-organ function and hemodynamics over preoperative values. Renal failure historically has been associated with mortality rates of 70% to 100% for patients undergoing cardiac operations [18]. Previous investigations with pneumatic ventricular assist devices have described the occurrence of renal failure to be highly predictive of mortality, and renal failure (as indicated by a serum creatinine level greater than two times normal values) itself was a contraindication to LVAD implantation in our study [19]. We report a marked improvement in renal function after the commencement of LVAD support, a finding consistent with our previous investigations [4]. These data indicate that mechanical circulatory assistance and subsequent improved renal perfusion may benefit the recipient's renal function before transplantation. Interestingly, after transplantation, creatinine values have been demonstrated to increase slightly, most likely due to the employment of potentially nephrotoxic immunosuppressive therapy. Hepatic failure (as indicated by serum transaminase levels greater than two times the normal values) is considered a contraindication to LVAD implantation because of the attendant risk of coagulopathy associated with hepatic dysfunction. We report a marked improvement in liver function with LVAD support as evidenced by serial serum transaminase samples. This beneficial effect was maintained for as long as 10 months, a finding consistent with previous reports [20]. Thus, although hepatic and renal failure per se before implantation may be regarded as potential contraindications to implantation, both demonstrate an excellent response to mechanical circulatory assistance. Implantation of ventricular assist devices also has been shown to reduce serum aldosterone levels, plasma renin activity, and plasma levels of atrial naturetic peptide [21]. In the current study, we report similar results in our LVAD patient cohort. These changes are likely a response to the improvement in blood flow. However, because the pathophysiologic consequences of heart failure are complex, such neurohormonal changes warrant further investigation as they may themselves have independent effects on renal, hepatic, and gastrointestinal perfusion. Finally, we retrospectively evaluated our cohort of patients for adverse clinical and mechanical events that potentially would have precluded their successful temporary outpatient discharge. The parameters used for this analysis represented the most common complications reported in the literature in conjunction with ventricular assistance: right heart failure, arrhythmias, thromboembolic and infectious complications, hemolysis and bleeding, end-organ dysfunction, and mechanical malfunction. Notably, most adverse clinical and mechanical events in our cohort occurred during the 30-day perioperative period, the majority of which involved infections, arrhythmias, and bleeding. There were no thromboembolic complications in any of these patients, the majority of whom received no anticoagulant therapy. After the perioperative period, however, the overall number of complications were few. Further, those adverse events that historically

LEVIN ET AL LEFT VENTRICULAR DEVICE AS OUTPATIENT THERAPY

1519

have resulted in significant morbidity-hemolysis, bleeding, and thromboembolic events-were absent in our cohort. Importantly, the perioperative period during which these adverse clinical and mechanical events occurred also represents a period during which hospital discharge would otherwise be unlikely. Those events that potentially would have precluded discharge and that occurred after the perioperative period involved infections (60% of which were associated with line sepsis) and arrhythmias. None of these events would have precluded safe return to the hospital or physician's office for treatment. Although the data do not indicate that no problems will occur outside of the hospital, they do support the notion that, for most candidates, even those with a history of arrhythmias or in whom infectious episodes have been the source of significant morbidity, outpatient discharge may be a feasible alternative to prolonged inpatient stays. Of note, although the pneumatic HeartMate 1000 IP LVAD evaluated in this study is not suitable for outpatient use, a portable pneumatic unit is now available that is functionally identical to the original device. The Institute of Medicine has predicted that as many as 50,000 patients will need some form of cardiac assistance [22]. Because the number of organ donors is unlikely to increase substantially so as to meet this need, it is clear that alternatives to human cardiac transplantation are needed urgently. It has been argued that the use of LVADs does not increase overall wait list survival but merely changes the order in which patients receive a donor organ. However, this outcome is no different from a change in priority due to the commencement of inotropic therapy. Although LVADs to date have been used only as a bridge to transplantation, investigations involving the outpatient experience of bridged patients will yield valuable information regarding the development of the device as a complete alternative to cardiac transplantation. In summary, these data indicate that functional and physiologic recovery occurs during LVAD support and that adverse clinical and mechanical events occurring after the perioperative period are few and should not preclude hospital discharge in selected patients. Thus, LVADs can provide reliable, longterm support and should be evaluated for future use as outpatient therapy. Supported in part by a grant from the Lowy Foundation.

References 1. Smith WM. Epidemiology of congestive heart failure. Am J

Cardiol 1985;55:3A. 2. United Network For Organ Sharing. UNOS update 1991;7:2. 3. Copeland JG, Emery RW, Levinson MM, Copeland J, McAleer MJ, Riley JE. The role of mechanical support and transplantation in treatment of patients with end stage cardiomyopathy. Circulation 1985;72(SuppI2):7-12. 4. Levin HR, Chen JM, Dasse KA, Graham TR. End organ function during mechanical left ventricular assistance. In: Mechanical circulatory assistance. Kent, England: Hodder and Stoughton (in press).

1520

Ann Thome Surg 1994;58:1515-20

LEVIN ET AL LEFf VENTRICULAR DEVICE AS OUTPATIENT THERAPY

5. Frazier OH, Rose EA, Macmanus Q, et al. Multicenter clinical evaluation of HeartMate 1000 IP left ventricular assist device. Ann Thorac Surg 1992;53:1080-90. 6. Kanter KR, McBride LR, Pennington DG, et al. Bridging to cardiac transplantation with pulsatile ventricular assist devices. Ann Thorac Surg 1988;46:134-40. 7. Dew MA, Kormos RL, Roth LH, et al. Life quality in the era of bridging to cardiac transplantation. ASAIO J 1993;39:145-52. 8. Bourassa MG, Gurne 0, Bangdiwala SI, et al. Natural history and patterns of current practice in heart failure. J Am Coll Cardiol 1993;22(Suppl A):14A-9A. 9. Massie BM, Packer M. Congestive heart failure: current controversies and future prospects. Am J Cardiol 1990;66:429-30. 10. National Center for Health Statistics. Vital Statistics of the United States 1988. Vol II mortality, Part A. Washington DC: Public Health Service 1991; DHHS publication no. (PHS) 91-1101. 11. Gillum RF. Heart failure in the United States 1970-1985. Am Heart J 1987;113:1043-5. 12. Graves EJ. Detailed diagnoses and procedures, National Hospital Discharge Survey, 1989. Hyattsville, MD: National Center for Health Statistics, 1991; DHHS publication no. (PHS) 91-1769 (Vital and Health Statistics; series 13; no. 108). 13. Ghali JK, Cooper R, Ford E. Trends in hospitalization rates for heart failure in the United States, 1973-1986: evidence for increasing population prevalence. Arch Intern Med 1990;150: 769-73. 14. Ho KK, Pinsky JL, Kannel WB, Levy D. Part II: New insights into the epidemiology and pathophysiology of heart failure

15. 16.

17. 18. 19. 20.

21. 22.

(Pitt B, Chairman). J Am Coll Cardiol 1993;22(Suppl A):16A23A. Yusuf S, Thorn T, Abbott RD. Changes in hypertension treatment and in congestive heart failure mortality in the United States. Hypertension 1989;13(Suppl 1):74-9. Pfeffer MA, Braunwald E, Moye LA, et al. Effect of captopil on mortality and morbidity in patients with left ventricular dysfunction after myocardial infarction: results of the Survival and Ventricular Enlargement Trial. N Engl J Med 1992;327: 699-77. Cooley DA, Liotta D, Hallman GL, Bloodwell RD, Leachman RD, Milam JD. Orthotopic cardiac prosthesis for two-staged cardiac replacement. Am J Cardiol 1969;24:723-30. Yeboah ED, Petrie A, Pead JL. Acute renal failure and open heart surgery. Br J Med 1972;1:415-8. Kanter KR, Swartz MT, Pennington DG, et al. Renal failure in patients with ventricular assist devices. ASAro Trans 1987;33: 426-8. Dasse KA, Frazier OR Lesniak JM, Myers T, Burnett CM, Poirier VL. Clinical responses to ventricular assistance versus transplantation in a series of bridge to transplant patients. ASAIO Trans 1992;38:M622-6. Quintero TE, Uretsky BF, Murali S, et al. Amelioration of the heart failure state with left ventricular assist system support. J Am Coll Cardiol 1992;19:254A. Hodgness JR, VanAntwerp M, ed. Executive summary. In: The artificial heart prototypes, policies and patients. Washington: National Academy Press, 1991:1-13.