Target Heart Failure Populations for Newer Therapies
Background The scarcity of donor hearts has created a large population of heart failure patients who are unlikely to undergo transplantation. Newer surgical therapies that might sustain such patients at home previously have been applied in critical situations in which early outcome is jeopardized by multiorgan failure. The optimal population for studies of extended support would be ambulatory patients with low operative risk but high risk of later unfavorable outcome.
Methods and Results Baseline clinical, echocardiographic, and hemodynamic data were collected prospectively between 1988 and 1993 in 500 patients who were discharged on tailored medical therapy after evaluation for transplantation. Specific criteria were examined to identify high risk of death or need for urgent transplantation during the next 2 years. In 265 patients with ejection fraction ≤25% and initial New York Heart Association class IV symptoms, survival at 2 years was 55% (without urgent transplantation, 45%). Lower cardiac index or higher filling pressures at the time of referral did not confer higher risk, which was predicted by persistence of higher pressures after therapy. Serum sodium below 133 was associated with 34% 2-year survival without urgent transplantation, and ventricular dimension >80 mm with a rate of 25%. Patients with initial peak oxygen consumption >10 mL/kg per minute had a 2-year event-free rate of 72% compared with 48% for those with <10 mL/kg per minute and 32% for those unable to exercise at referral. Demonstration of a 30% decrease in mortality with a controlled trial of new therapy in patients with ejection fraction ≤25% would require 600 patients with class III symptoms or almost 300 patients with class IV symptoms unless another criterion were added.
Conclusions Ambulatory populations with high predicted event rates can be identified at initial evaluation, when hemodynamic criteria may be less useful than ventricular dimension, serum sodium, and ability to exercise. The use of outcome data from previous eras may lead to overestimation of benefits from newer therapies and underestimation of the sample size required in a prospective trials.
The promise of transplantation has encouraged the identification and referral of patients with heart failure previously considered to be terminal. The shortage of donor hearts and the complications of immunosuppression and rejection limit the benefits of transplantation to a select minority of patients with advanced heart failure.1 Newer surgical therapies such as insertion of mechanical support devices have been reserved until patients are hospitalized with severe decompensation from which multiorgan dysfunction often hinders recovery.2 3 Approaches using skeletal muscle have focused on the ambulatory population without severe decompensation.4 The optimal population for extended support would be ambulatory patients who have low operative risk and yet have high risk of unfavorable outcome over the next 2 years.
Efficient design of prospective trials requires estimation of the current “natural history” for groups of heart failure patients, which is influenced by many factors. The expertise and intensity of follow-up care for posttransplant patients have in some centers been extended to the medical therapy of potential transplant candidates as well. The availability of transplantation also changes survival rates. Because cardiac transplantation is the best current therapy for eligible patients with truly refractory decompensation, in whom death would otherwise be imminent, such patients cannot be merely “censored” from survival analyses of other therapies. It is also misleading, however, to infer retrospectively that transplantation was necessary merely because it was performed, particularly in the absence of uniform adherence to selection criteria such as those approved by the Bethesda Conference.5
Many risk factors for mortality with mild to moderate heart failure have been identified.6 7 Only those that are relevant and easily applied in advanced heart failure could be used to define patient populations for multicenter studies of newer therapies that carry significant early risk. The purposes of this study were (1) to determine from data acquired prospectively the current frequency of the combined end point of mortality and urgent transplantation for patients followed in a specialized heart failure/transplant center and (2) to identify broad parameters of risk that are relatively independent of therapeutic vigor and could be applied uniformly at multiple centers. These parameters ideally should define a high-risk study group of adequate size for controlled trials of current investigational therapies.
Baseline information including clinical, echocardiographic, and hemodynamic profiles was collected and recorded prospectively on all patients discharged after hospitalization and complete evaluation for cardiac transplantation at the Ahmanson–University of California, Los Angeles, Cardiomyopathy Center. To analyze outcomes in the current era of therapy, analysis was restricted to patients evaluated between January 1988 and March 1993. Patients considered to have New York Heart Association class I or II symptoms at referral were excluded from this analysis, except for 43 patients added to one analysis of left ventricular dimension and ejection fraction ≤25%. Because the large majority of patients in this relatively young population had systolic failure, only patients with ejection fractions ≤25% were included.
Echocardiography was performed within 1 week of initial hospitalization. Left ventricular ejection fraction was measured with the use of a modified Simpson’s rule. Left ventricular end-diastolic dimension was measured just beyond the level of the mitral leaflets. Right heart catheterization was performed as part of the transplant evaluation. The hemodynamics obtained at this time were considered “initial.”
Exercise gas analysis was performed with the use of the Medical Graphics 2000 metabolic cart and a 15-W ramp bicycle protocol in patients who were ambulatory at the time of initial evaluation without symptomatic congestion at rest or severe angina. Only exercise testing during initial evaluation is included in this data analysis.
Therapy was adjusted during hemodynamic monitoring for patients with initial pulmonary capillary wedge pressure >20 mm Hg or cardiac index <2.2 L · min−1 · m−2. Vasodilators and diuretic agents were tailored to approach hemodynamic goals of pulmonary capillary wedge pressure ≤15 mm Hg and systemic vascular resistance ≤1200 dyne · s · cm−5 (in the average-sized individual) while maintaining systolic blood pressure ≥80 mm Hg, as previously described for this population.8 9 From 1988 to mid-1990, vasodilator therapy included either captopril or hydralazine in combination with oral nitrate therapy; after the results of the Hy-C Trial in October 1990,10 all patients initially received angiotensin-converting enzyme inhibitors and isosorbide dinitrate if tolerated. Hydralazine was added if necessary to achieve adequate vasodilation without unacceptable side effects. In these patients, the hemodynamic parameters recorded on the adjusted oral regimen also were recorded and analyzed for predictive value.
The flexible diuretic regimen of loop diuretics was based on daily weights, with intermittent supplementation with metolazone if necessary. Freedom from signs or symptoms of congestion was maintained in the majority of patients. Amiodarone was the preferred antiarrhythmic agent if therapy was considered to be indicated for atrial or ventricular arrhythmias. Patients subsequently were followed in the Cardiomyopathy Center in collaboration with their referring cardiologists.
Candidacy for Transplantation
Indications for cardiac transplantation followed the general guidelines now accepted.5 Absolute and relative contraindications were considered according to standard practice.5 11 Patients in whom relative contraindications such as noncompliance led to initial rejection with subsequent chance for reevaluation were rejected at initial evaluation for this analysis. Patients were hospitalized in an intensive care unit for further therapy with intravenous inotropic agents or mechanical assistance when adequate systemic perfusion and noncardiac organ function could not be maintained otherwise or when intractable ventricular arrhythmias or clinical ischemia could not be controlled otherwise. Persistence of these conditions in the intensive care unit was considered an indication for urgent transplantation in eligible candidates. The waiting period for urgent transplantation varied from 1 to 90 days during this period.
Outcomes were determined as of April 1994, at which time all surviving patients had a minimum of 1 year of follow-up. Only 2.5% of patients were lost to follow-up. Sudden death was defined as death occurring out of the hospital instantaneously, within 15 minutes of a change in symptoms, or unexpectedly during sleep. Hospitalizations were recorded as end points when they included transplantation.
All statistical analyses were performed with the bdmp statistical package.12 The majority of analyses were performed in patients with both ejection fraction ≤25% and initial symptoms considered to be New York Heart Association class IV. Actuarial survival rates and actuarial survival rates without urgent transplantation were determined by the Kaplan-Meier product limit estimate. Comparison between two groups was performed with the Mantel-Cox statistic. Because of the small numbers of patients with greater than 2-year follow-up, the survival analyses were truncated at 730 days. Comparison of multiple groups defined by values of serum sodium or left ventricular dimension was performed with the Mantel-Cox and Breslow statistics for trend.
The Cox multivariate analysis was performed with the use of variables of initial and final filling pressures and cardiac indexes, presence of coronary artery disease, echocardiographic ejection fraction, left ventricular diastolic dimension, and serum sodium. Exercise performance was examined but was available in only 107 patients. The median values for hemodynamic parameters before and after hemodynamically guided therapy in New York Heart Association class IV patients were used to probe for groups with greater risk than that conferred by the ejection fraction ≤25% and class IV clinical class. The parameters of left ventricular dimension and serum sodium were studied in the most detail. Therapy with angiotensin-converting enzyme inhibitors and amiodarone also were included as variables, although the selection of therapies was influenced by other clinical parameters.
Sample-size calculations were performed according to the method of Freedman.13 Trial design was based on the assumption that a new therapy might decrease mortality by 30% with an α error of 5% and a 1−β of 80%. The two-tailed t test was used. All means are expressed as ±SD.
Patients With New York Heart Association Class IV Symptoms and Ejection Fraction <25%
The survival rate for the 404 patients (1988 to 1993) discharged after referral for transplantation with New York Heart Association class III or IV symptoms and left ventricular ejection fraction ≤25% was 74% at 1 year and 61% at 2 years. Survival without urgent transplantation was 67% at 1 year and 53% at 2 years, respectively (Fig 1⇓). Including only patients with ejection fraction ≤20% reduced the population size to 250, in whom survival without transplantation was reduced only to 63% and 49% at 1 and 2 years, respectively. Although threshold ejection fractions of 30% and 25% were little different, the value of 25% was chosen as a more typical descriptor of patients evaluated with advanced heart failure.
For the 265 patients with ejection fraction ≤25% and New York Heart Association class IV symptoms, the actuarial survival rate was 68% at 1 year and 55% at 2 years (Fig 2⇓). Including the likelihood of urgent transplantation (29 patients) over the 2 years, the overall freedom from death or urgent transplantation was 59% and 45% at 1 and 2 years, respectively. Sudden death accounted for 37% of mortality. These numbers are compared with published results for class IV heart failure in previous eras (Fig 2⇓14 15 ).
The majority of patients were men, and coronary disease was implicated in ≈50% of the patients (Table 1⇓). Therapy at the time of discharge included angiotensin-converting enzyme inhibitors in 61% of patients and amiodarone in 47% of patients. After 1990, absence of therapy with angiotensin-converting enzyme inhibitors indicated intolerance to these agents, which is more likely in the setting of severe decompensation.10 The changes between initial hemodynamics and hemodynamics on the discharge regimen reflect the impact of therapy tailored to reduce ventricular filling pressures and systemic vascular resistance.
As a result of initial evaluation, 143 of the 265 (54%) patients with ejection fraction ≤25% and initial New York Heart Association class IV symptoms were found acceptable for transplantation, although actual listing was frequently deferred or inactivated as patients improved. Outpatient transplantation, which was a censored event, was performed in 42 patients during the first year after initial evaluation and in another 4 patients during the second year. For the 122 patients found to have major contraindications during initial evaluation, actuarial survival rates were 51% at 1 year and 37% at 2 years, compared with 65% at 1 year and 56% at 2 years without urgent transplantation in potentially acceptable candidates (P=.004).
Multivariate Analysis of Clinical Profiles
Although this study was not designed to be another multivariate analysis of all possible prognostic factors, multivariate proportional hazards analysis was performed to confirm the importance of variables selected for further analysis. Including ejection fraction, coronary artery disease, serum sodium, left ventricular dimension, and initial and treated cardiac indexes, pulmonary capillary wedge pressures, and right atrial pressures, analysis yielded pulmonary capillary wedge pressure on therapy, right atrial pressure on therapy, left ventricular diastolic dimension, and serum sodium as predictive of outcome (Table 2⇓).
Left ventricular ejection fraction was of borderline significance in some analyses (the exact composition of the subjects varied as a result of occasional missing data such as unmeasurable left ventricular dimensions) but was not an independent predictor in any analysis. Although coronary artery disease has been a predictor in broader populations, in this population, defined by initial New York Heart Association class IV symptoms and ejection fraction ≤25%, survival without urgent transplantation at 1 year was 57% with and 61% without coronary artery disease and at 2 years was 42% versus 48% with and without coronary artery disease, respectively (P=.21). Sudden death–free survival also was similar: 86% at 1 year with coronary artery disease compared with 87% without. The presence of a history either of syncope or cardiac arrest, both shown to predict poor outcome in a wider population of heart failure, was not associated with higher mortality in univariate or multivariate proportional hazards analysis; the sudden death survival rate in patients with a positive history was 82% at 1 year compared with 88% without such a history (P=.3). Further analysis of sudden death was not done because of the limited number of sudden death end points.
Limitations of Hemodynamic Criteria
Hemodynamic criteria were examined closely, as they traditionally have been used to define severity of disease and indications for support devices and other advanced therapies. Because there were large changes between hemodynamics at referral and hemodynamics on the tailored oral regimen before discharge (Table 3⇓), both sets were considered. The groups were first divided at the median values for initial cardiac index, pulmonary capillary wedge pressure, and right atrial pressure. Comparing the 2-year actuarial curves for groups above and below the initial medians showed values worse than the mean to confer no increase in the chance of adverse outcome (Fig 3⇓).
After undergoing therapy directed to minimize filling pressures, patients with pulmonary capillary wedge pressure above the median treated value of 16 mm Hg had a 38% 2-year survival rate without urgent transplant compared with 50% for lower pulmonary capillary wedge pressure (P=.007). A similar difference was conferred by the lower right atrial pressure. Use of either of these criteria for population selection increased the frequency of end points but not by a large amount (Fig 4⇓). The survival rate without urgent transplantation was 45% at 2 years for patients whose cardiac index on revised therapy was above the median 2.5 L · min−1 · m−2 and the same for those whose treated cardiac index on revised therapy was below the median.
Serum Sodium and Left Ventricular Dimension
Survival for New York Heart Association class IV patients with ejection fraction ≤25% was determined for patients using different threshold values for serum sodium. Because the purpose was to define the largest possible group with high risk, the survival data shown in Table 4⇓ indicate the worsening actuarial survival rates obtained by eliminating successively more patients on the basis of higher serum sodium concentration at the time of hospital discharge. The strictest criterion for sodium, <130 mEq/L, identified patients with a 2-year survival rate without urgent transplantation of only 33%. This criterion, however, restricted the population to only 46 patients, while using a threshold serum sodium of 134 mEq/L included 116 patients with only slightly higher chance of favorable 2-year outcome (37%).
Similar analysis of the left ventricular end-diastolic dimension from echocardiography could be done for 240 of the patients with initial New York Heart Association class IV symptoms and ejection fraction ≤25% (Table 5⇓). Including only those with dimension ≥80 mm identified a group of 78 patients with likelihood of 2-year survival without urgent transplantation of 25%.
Because the optimal criteria for broad application are those requiring no subjective clinical assessment, the analysis was repeated excluding all classification of New York Heart Association class, which was included in the above analyses. The two echocardiographic criteria of left ventricular ejection fraction ≤25% and left ventricular diastolic dimension ≥80 mm Hg selected 128 patients for whom survival without urgent transplantation was 56% at 1 year and 34% at 2 years (Fig 5⇓). In this group 31 of the 65 deaths occurred suddenly at home and 17 patients underwent urgent transplantation.
Exercise Testing in Class IV Patients
For the 107 patients who performed exercise testing at the time of presentation with New York Heart Association class IV symptoms and ejection fraction ≤25%, peak oxygen consumption was not predictive of outcome by multivariate analysis. This most likely reflects the small number of tests and the selection bias against exercise in patients with resting symptoms. If the mere performance of an exercise test were included in multivariate analysis, it was an independent predictor of good outcome in this compromised population. Among the exercising patients, the 26 with peak |$$˙o2 ≤10 mL/kg per minute had a 2-year rate of survival without urgent transplantation of 48% compared with 72% (P=.03) for all patients with higher initial peak |$$˙o2, with rates of 76%, 78%, and 75% for peak |$$˙o2 10 to 12, 12 to 14, and 14 to 16 mL/kg per minute. Patients not exercising had a 32% rate of survival at 2 years without urgent transplantation (Fig 6⇓).
Influence of Specific Therapies
The period of this study encompassed evolving therapies, which were given according to best clinical judgment at the time, or in some cases, specific randomized protocols.10 Therapy with angiotensin-converting enzyme inhibitors, which is more likely to be tolerated in patients with less severe hemodynamic compromise,10 was an independent predictor of survival. Therapy with amiodarone, given for symptomatic ventricular arrhythmias, atrial fibrillation, or frequent high-grade ventricular ectopy, also was found to confer survival benefit despite the possible adverse prognostic impact of the indications for therapy with this agent (Table 2⇑).
Prognosis and Trials in New York Heart Association Class IV Heart Failure
This analysis of patients discharged after referral for heart transplantation between 1988 and 1993 confirms the improving survival rate for patients presenting with New York Heart Association class IV heart failure.14 15 Therapeutic advances over the past 10 years include the broader use of angiotensin-converting enzyme inhibitors,10 16 17 avoidance of type I antiarrhythmic agents18 19 in favor of amiodarone, and aggressive reduction of elevated filling pressures.8 9 20 The expected survival rate on medical therapy at 2 years for class IV heart failure and ejection fraction ≤25%, however, remains <50% when urgent transplantation is included as a true end point for failure of medical therapy. The survival rate with heart transplantation is much higher, currently 70% to 80% at 2 years.21
This grim outlook without transplantation confronts the 40% to 70% of patients who have advanced heart failure but are found to be ineligible at major centers. In addition, the scarcity of donor hearts has created long lists for outpatient candidates, who face considerable risks during the wait of over 18 months for transplantation.1 This wait is predicted to increase even further during the next 4 years, after which there may be little chance of transplantation for any outpatient candidate if current trends continue.
From these ineligible and waiting groups of ambulatory patients, it should be possible to select target groups for alternative interventions such as mechanical devices, skeletal muscle assistance, or experimental drug therapy. The true impact of these interventions will be better assessed when applied in patients who have not deteriorated to multiorgan failure that compromises early outcomes. Because controlled trials in these challenging patients require extensive adjunctive care and heavy commitment of resources, both the relative risks and costs could be minimized by identifying an ambulatory population sufficiently large to ensure enrollment but at sufficiently high risk with current therapy to ensure adequate end points for comparison. The present comprehensive analysis of initial clinical-hemodynamic profiles and subsequent outcomes in patients referred to a heart failure–transplant center is the largest currently available from which to design a trial in ambulatory patients with advanced heart failure.
Initial hemodynamic criteria have been used in some previous trials of mechanical support2 3 as appropriate for patients in intensive care. Among the patients who could be rendered ambulatory in this study, however, initial hemodynamic parameters, defined as the median group values, would restrict the size of the population without increasing the event rate. The vigor of recent therapy is the main determinant of initial filling pressures and cardiac index in patients presenting with severe symptoms. After therapy adjusted to minimize filling pressures and reduce systemic vascular resistance, persistence of elevated filling pressures did confer significantly worse prognoses. The importance of minimal filling pressures has been shown previously for the population of patients referred for transplantation with ejection fractions ≤20%22 and more recently for the 660 patients discharged on vasodilator therapy after referral to transplantation.23 Like the current analysis, that study showed cardiac index to be of no value in prognosis. It should be emphasized that cardiac index may be predictive in other populations in which few patients present with New York Heart Association class IV symptoms.7
Other Potential Criteria for Trial Design
The lability and lack of specificity of hemodynamic parameters make them difficult criteria to use in multicenter trials of patients considered to be New York Heart Association class IV. The alternative prognostic factors selected for further analysis were those identified from previous eras in this center and others.24 25 26 Lower serum sodium, as a reflection of neurohormonal activation and the net effect of stimuli for fluid retention, was helpful in describing this population. Excluding patients with serum sodium values >136 would reduce the population by not quite half, with an expected increase in event rate. Strict exclusion to serum sodium values <130 would create only a very small population without a large increase in event rate. Such specific selection may be appropriate only for therapies specifically targeted to this neurohormonal activation, as described previously for hyponatremia and angiotensin-converting enzyme inhibition.24
Peak oxygen consumption has become an integral part of evaluation for transplantation, particularly as more standard indications are sought.5 26 In this population of patients presenting with New York Heart Association class IV symptoms of heart failure, however, the majority of patients were not considered sufficiently stable to undergo testing during initial evaluation. In those who did, only the threshold value of 10 mL/kg per minute identified a higher risk group. Standard use of exercise testing immediately before hospital discharge might be a consistent way for multiple centers to identify patients with comparable functional limitation. Testing at any one time, however, would be strongly influenced by the period of prior bed rest and deconditioning.
Massive left ventricular dilation was associated with higher risk of sudden death and progression to heart failure death or urgent transplantation, as has been shown in analysis of an earlier population at this center.25 Left ventricular dilation leads to greater wall stress, myocardial oxygen demands, heterogeneity of repolarization, and activation of intracardiac reflexes. This deleterious effect appeared to be independent of direct hemodynamic correlates.25
Left ventricular dimension was previously found to be relatively stable over weeks on therapy, unlike many parameters of heart failure.25 This stability, combined with the possible objectivity of measurement, makes left ventricular dimension a desirable criterion for selection of high-risk patients from multiple centers. Eliminating even the subjective New York Heart Association classification and using only the echocardiographic parameters of left ventricular ejection fraction ≤25% and left ventricular diastolic dimension ≥80 mm identified 128 ambulatory patients, of whom only 34% would be predicted to survive 2 years without urgent transplantation.
The outcome data from this population may facilitate design of controlled trials to test these newer therapies against optimal medical therapy.13 To test a newer therapy reducing mortality by 30% in a controlled trial of class III patients on optimal medical therapy could require a sample size of 660 patients and accepting an α error of 5% and a β error of 20%, with a two-tailed test (Table 6⇓). To test a therapy with similar potential for New York Heart Association class IV patients with ejection fraction ≤25% and left ventricular diastolic dimension >78 mm would require 170 patients.
Left ventricular ejection fraction was measured from echocardiography, which may differ from other techniques but is the method most commonly used for screening of this population. The importance of peak oxygen consumption in this compromised population cannot be assessed easily because of the number of patients too compromised to undergo testing. This analysis is necessarily retrospective. All baseline data and outcome data were collected prospectively, however, in a database that included all patients undergoing evaluation. The specific factors found to influence outcome have been identified previously by other investigators and in analyses of years before 1988 in the experience in this center. Extrapolation of these results to other experiences may be limited by varying intensities of medical therapy between centers. In addition, referral biases related to perceptions about transplantation affect the complexion of a heart failure population at a transplant center. In a primary care setting, some patients with similar objective criteria may have less clinical compromise. In addition, almost all patients were younger than 70 years, while older patients may be considered for procedures other than transplantation.
This analysis of an advanced heart failure population provides guidance for the design of future trials in which high early event rates should allow assessment of alternative therapies that carry their own risks.27 While the specific criteria suggested here include ejection fraction, serum sodium, and left ventricular dimension, the results emphasize the general importance of choosing inclusion criteria specific for each desired study population rather than those generalized from other populations and historical therapies. In addition, the frequency of urgent transplantation as an end point in advanced heart failure mandates careful consideration regarding inclusion of potential transplant candidates in a randomized study.
- Copyright © 1995 by American Heart Association
Stevenson LW, Warner SL, Steimle AE, Fonarow GC, Hamilton MA, Moriguchi JD, Kobashigawa JA, Tillisch JH, Drinkwater DC, Laks H. The impending crisis awaiting cardiac transplantation: modeling a solution based on selection. Circulation. 1994;89:450-457.
Cohn JN, Johnson GR, Shabetai R, Loeb H, Tristani F, Rector T, Smith R, Fletcher R. Ejection fraction, peak exercise oxygen consumption, cardiothoracic ratio, ventricular arrhythmias, and plasma norepinephrine as determinants of prognosis in heart failure. Circulation. 1993;87(suppl VI):VI-6-VI-16.
Stevenson LW, Tillisch JH. Maintenance of cardiac output with normal filling pressures in dilated heart failure. Circulation. 1986;74:1303-1308.
Fonarow GC, Chelimsky-Fallick C, Stevenson LW, Luu M, Hamilton MA, Moriguchi JD, Tillisch JH, Creaser JA, Albanese E. Effect of direct vasodilation vs angiotensin-converting enzyme inhibition on mortality in advanced heart failure: the Hy-C Trial. J Am Coll Cardiol. 1992;19:842-850.
Dixon WJ, ed. In: bmdp Statistical Software Manual. Los Angeles, Calif: University of California Press; 1992.
Woolson RF. Statistical Methods for the Analysis of Biomedical Data. New York, NY: Wiley & Sons; 1987:457-459.
Cohn JN, Johnson G, Ziesche S, Cobb F, Francis G, Tristani F, Smith R, Dunkman WB, Loeb H, Wong M, Bhat G, Goldman S, Fletcher RD, Doherty J, Highes CV, Carson P, Cintron G, Shabetai R, Haakenson C. A comparison of enalapril with hydralazine-isosorbide dinitrate in the treatment of chronic congestive heart failure. N Engl J Med. 1991;325:303-310.
Stevenson WG, Stevenson LW, Middlekauff HR, Saxon LA. Sudden death prevention in patients with advanced ventricular dysfunction. Circulation. 1993;88:2953-2961.
Stevenson WG, Stevenson LW, Middlekauff HR, Fonarow GC, Woo MA, Saxon LA, Natterson PD, Tillisch JH. Results of major trials influence therapy and survival for patients with advanced heart failure. Circulation. 1994;90(suppl I):I-380. Abstract.
Stevenson LW, Tillisch JH, Hamilton M, Luu M, Chelimsky-Fallick C, Moriguchi JD, Kobashigawa J, Walden J. Importance of hemodynamic response to therapy in predicting survival with ejection fraction ≤20% secondary to ischemic or nonischemic dilated cardiomyopathy. Am J Cardiol. 1990;66:1348-1354.
Stevenson LW, Fonarow G, Hamilton M, Tillisch JH. Why cardiac output is not a good hemodynamic target for therapy in advanced heart failure. Circulation. 1994;90(suppl I):I-611. Abstract.
Lee WH, Packer M. Prognostic importance of serum sodium concentration and its modification by converting enzyme inhibition in patients with severe chronic heart failure. Circulation. 1986;73:257-267.
Mancini DM, Eisen H, Kussmaul W, Mull R, Edmunds LH Jr, Wilson JR. Value of peak oxygen consumption for optimal timing of cardiac transplantation in ambulatory patients with heart failure. Circulation. 1991;83:778-783.