Author + information
- Received June 3, 2003
- Revision received August 23, 2003
- Accepted August 26, 2003
- Published online March 3, 2004.
- Javed Butler, MD, MPH, FACC*,†,‡∥,¶,* (, )
- Ghazanfar Khadim, MD*,
- Kimberly M. Paul, MD§,
- Stacy F. Davis, MD*,
- Marvin W. Kronenberg, MD*,
- Don B. Chomsky, MD*,¶,
- Richard N. Pierson III, MD# and
- John R. Wilson, MD*
- ↵*Reprint requests and correspondence:
Dr. Javed Butler, Cardiology Division, 383 PRB, Vanderbilt University Medical Center, Nashville, Tennessee 37232-6300, USA.
Objectives We sought to assess the relationship between survival, peak exercise oxygen consumption (Vo2), and heart failure survival score (HFSS) in the current era of heart failure (HF) therapy.
Background Based on predicted survival, HF patients with peak Vo2<14 ml/min/kg or medium- to high-risk HFSS are currently considered eligible for heart transplantation. However, these criteria were developed before the widespread use of beta-blockers, spironolactone, and defibrillators—interventions known to improve the survival of HF patients.
Methods Peak Vo2and HFSS were assessed in 320 patients followed from 1994 to 1997 (past era) and in 187 patients followed from 1999 to 2001 (current era). Outcomes were compared between these two groups of patients and those who underwent heart transplantation from 1993 to 2000.
Results Survival in the past era was 78% at one year and 67% at two years, as compared with 88% and 79%, respectively, in the current era (both p < 0.01). One-year event-free survival (without urgent transplantation or left ventricular assist device) was improved in the current era, regardless of initial peak Vo2: 64% vs. 48% for peak Vo2<10 ml/min/kg (p = 0.09), 81% vs. 70% for 10 to 14 ml/min/kg (p = 0.05), and 93% vs. 82% for >14 ml/min/kg (p = 0.04). Of the patients with peak Vo2of 10 to 14 ml/min/kg, 55% had low-risk HFSS and exhibited 88% one-year event-free survival. One-year survival after transplantation was 88%, which is similar to the 85% rate reported by the United Network for Organ Sharing for 1999 to 2000.
Conclusions Survival for HF patients in the current era has improved significantly, necessitating re-evaluation of the listing criteria for heart transplantation.
Heart failure (HF) patients with peak exercise oxygen consumption (Vo2) of <14 ml/min/kg or a medium- to high-risk heart failure survival score (HFSS) are currently considered eligible for heart transplantation, based on the presumption that such patients will have a better survival with transplantation than with medical therapy (1–5). However, these criteria were developed before the widespread use of beta-blockers, spironolactone, and defibrillators—interventions known to improve survival in HF (6–11). These criteria may therefore need to be revised in the current era. To test this hypothesis, we compared the relationship between peak Vo2, HFSS, and survival between HF patients treated from 1994 to 1997 and those treated from 1999 to 2001, and we compared their survival with patients who underwent heart transplantation between 1993 and 2000.
The study population consisted of HF patients with an ejection fraction <40% who underwent exercise testing with peak Vo2determination between 1994 and 1997 (past era) and 1999 to 2001 (current era), as part of the Vanderbilt Heart Failure Program, and patients who underwent heart transplantation between 1993 and 2000. Patients were excluded if they were taking inotropes, had angina or orthopedic problems restricting exercise capacity, had significant valvular stenosis, or exhibited oxygen desaturation during exercise. A total of 320 patients from the past era, 187 patients from the current era, and 184 patients who underwent transplantation were studied. Heart failure patients within each era were classified a priori, based on peak Vo2, into three groups: group 1 = <10 ml/min/kg; group 2 = 10 to 14 ml/min/kg; and group 3 = >14 ml/min/kg. Similarly, the HFSS was calculated for all patients, as described in the original investigation, using the following variables: etiology, heart rate, ejection fraction, mean blood pressure, intraventricular conduction delay, peak Vo2, and serum sodium (12). Based on the HFSS, these patients were classified into low- (>8.10), medium- (7.20 to 8.10), and high-risk (<7.19) groups.
Patients exercised on a Marquette treadmill, per the modified Naughton protocol, and were connected to a Medgraphics Cardio O2combined Vo2/ECG Exercise System (Medical Graphics Corp., St. Paul, Minnesota). The testing protocol and data interpretation have been previously described (13).
Outcomes and definitions
The primary study outcome was one-year event-free survival (without the need for a left ventricular assist device [LVAD] or urgent transplantation) for HF patients and overall one-year survival for transplanted patients. Urgent transplantation for the current era was defined as United Network for Organ Sharing (UNOS) status 1A listing at the time of transplantation. In the past era, urgent transplantation was defined as the presence, at the time of transplantation, of mechanical circulatory support with a left and/or right ventricular assist device implanted <30 days or evidence of complications related to mechanical support, intra-aortic balloon pump, mechanical ventilation, or continuous infusion of high-dose intravenous inotropes or multiple intravenous inotropes with continuous hemodynamic monitoring.
Univariate analyses were performed to assess associations between patient characteristics and outcomes, using the chi-square test for categorical variables and the ttest for continuous variables. The Kaplan-Meier method was used to assess survival. To assess the temporal trends in outcomes for patients undergoing transplantation, in order to match them with the two groups of HF patients, transplanted patients were also classified into two groups: those who were transplanted between 1993 and 1997 and those transplanted between 1998 and 2000. Because there was no significant survival difference between the two groups, they were combined for further outcome comparisons with the HF patients.
For the overall survival analysis for HF patients, all patients who underwent transplantation or LVAD placement were censored at the time of surgery. For event-free survival analysis, patients who underwent transplantation and did not meet the criteria for urgent transplantation were censored.
Exploratory analyses using the chi-square test were performed among the patients in the current era to assess the combined impact of beta-blocker and defibrillator therapy on one-year outcomes and the relationship between peak Vo2and HFSS. All analyses were performed using SPSS for Windows, release 11.5 (SPSS Inc., Chicago, Illinois).
Baseline patient characteristics of HF patients from the two eras are shown in Table 1. There were no significant differences between the two groups with respect to age, gender, race, ejection fraction, serum sodium, intraventricular conduction delay, and peak Vo2. Patients in the current era had, on average, a lower heart rate (74 ± 14 beats/min vs. 88 ± 16 beats/min, p < 0.01).
Baseline medical therapy was similar, except for higher use of beta-blockers (72% vs. 10%, p < 0.01) and spironolactone (41% vs. 2%, p < 0.01) in the current era. Beta-blocker types and dosages in the current era were as follows: 54 ± 33 mg atenolol (3%), 31 ± 24 mg carvedilol (54%), and metoprolol (43%)—108 ± 59 mg succinate and 64 ± 27 mg tartrate. Defibrillator use was also more common in the current era (19% vs. 11%, p = 0.01). The overall use of anti-arrhythmic medications was similar between the two eras (12% vs. 14% in the past vs. current era); however, for those with anti-arrhythmic therapy, there was a higher proportional use of amiodarone in the current era (82% vs. 70%, p = 0.01).
For patients who underwent heart transplantation, the mean age of the recipients at the time of transplantation was 52 ± 11 years. Of the patients, 81% were males, and 95% were white. The mean age of the donors was 29 ± 12 years. The average donor ischemic time was 181 ± 56 min.
The mean follow-up period for the past-era group was 311 ± 213 days; current era, 376 ± 268 days; and post-transplant group, 1,260 ± 854 days. Three patients in the past era and six in the current era underwent LVAD placement. A total of 52 patients in the past era and 22 in the current era died, and 40 patients who underwent transplantation died during the follow-up period. In the past era, 51 patients underwent heart transplantation, 12 of whom met the criteria for urgent transplantation. In the current era, 29 patients underwent heart transplantation, 8 of whom met the criteria for urgent transplantation.
Survival curves for the three groups are compared in Figure 1. Overall survival rates in the three groups (past era, current era, and those who underwent transplantation, respectively) were 78%, 88%, and 88% at one year and 67%, 79%, and 84% at two years (p < 0.01).
One-year event-free survival improved in the current era, regardless of initial peak Vo2: 64% versus 48% for peak Vo2of <10 ml/min/kg (p = 0.09), 81% versus 70% for peak Vo2of 10 to 14 ml/min/kg (p = 0.05), and 93% versus 82% for peak Vo2of >14 ml/min/kg (p = 0.04). When stratified by HFSS, the outcomes improved the most in the current era for the high-risk group (64% for current era vs. 41% for past era, p = 0.06) (Fig. 2).
Table 2shows the distribution of patients and outcomes based on both peak Vo2and HFSS. Of the patients with peak Vo2of 10 to 14 ml/min/kg, 55% had a low-risk HFSS and 88% one-year event-free survival. Similarly, 20% of patients with a medium-risk HFSS had peak Vo2of >14 ml/min/kg and demonstrated 85% one-year event-free survival.
Figure 3shows the event-free one-year outcomes for patients with intermediate-risk peak Vo2and a medium-risk HFSS, based on beta-blocker therapy and use of a defibrillator. Both groups had survival comparable to one-year survival after heart transplantation (88%) if the patients had a defibrillator and were taking beta-blockers.
The primary objective of heart transplantation is to improve survival and functional capacity. Consequently, patients are listed for transplantation only when it is presumed that they will live longer and function better after transplantation than they would with medical therapy. A variety of criteria has been developed in an attempt to identify such patients. The most widely utilized criteria are based on peak Vo2and HFSS, a scoring system based on seven parameters, including peak Vo2(12,14). Specifically, peak Vo2of <14 ml/min/kg or a medium- to high-risk HFSS is thought to identify patients who will have a survival benefit with heart transplantation (1–5). The validity of these criteria has not been reassessed since the initial recommendations. Recently, several new HF therapies have been introduced (e.g., beta-blockers, spironolactone, cardiac defibrillators), all of which improve survival (6–11). There continues to be a shortage of donor organs, leading to long waiting times and high mortality rates for patients awaiting transplantation (15,16). This necessitates that the available organs should be allocated to the patients at highest risk of dying, underscoring the importance of accurate prognosis determination.
Our data show that in the current era of HF management, the prognosis of patients has improved significantly, suggesting that the strategy for transplant candidate selection needs reassessment. Mancini et al. (14)demonstrated that HF patients with peak Vo2of <14 ml/min/kg have a one-year combined mortality or urgent transplantation rate of ∼50%, leading to a general recommendation that these patients should be evaluated for transplantation (1–5,14). Because peak Vo2is linearly related to HF prognosis, an alternate approach is to use it as continuous variable (17). Aaronson et al. (12)used this approach and incorporated other risk predictors to create a risk assessment model called the HFSS, which is now an alternate tool used to select heart transplant candidates. Because of these earlier reports, management of patients with advanced HF has undergone major changes that not only impact overall survival but also the relationship between peak Vo2or HFSS and survival. For example, beta-blockers improve survival but do not alter peak Vo2significantly (6). A benefit from defibrillators is also independent of changes in exercise tolerance. Moreover, beta-blockers also affect physiologic variables that constitute the HFSS (e.g., heart rate and mean blood pressure), thus modifying the HFSS calculation. It is therefore not surprising that beta-blocker therapy influences HF prognostic variables and that for every level of exercise tolerance, beta-blocker therapy may improve outcomes (18,19).
Similarly, our data indicate that in the current era of HF management, although peak Vo2continues to be an important predictor, the prognosis of patients with similar values has improved considerably, as compared with the past era. Even when LVAD and urgent transplantation were considered as “events,” along with death, one-year event-free survival was 81% for patients with peak Vo2of 10 to 14 ml/min/kg and increased to 86% for those patients who were on beta-blockers and had a defibrillator. These results are comparable to one-year post-transplant survival of 85%, on average, and demonstrate that for patients with intermediate-risk peak Vo2in the current era, medical therapy offers a survival benefit similar to heart transplantation (16).
If mortality rates are similar, should patients be transplanted to improve quality of life? Symptoms and the hospitalization rate improve significantly for HF patients treated with beta-blockers, whereas observational data suggest significant continued medical and psychological problems after transplantation (20–22). Considering the significantly improved morbidity and mortality of HF patients with medical therapy, it is not surprising that a transplant benefit is shown to be mostly restricted to the highest risk patients, and a randomized trial comparing medical therapy with transplantation has been suggested (23,24).
Current guidelines recommend that patients with peak Vo2of 10 to 14 ml/min/kg or a medium-risk HFSS should be considered for transplantation, thus tacitly equating the two groups of patients. However, there is a significant discordance between the two. Of the patients with peak Vo2of 10 to 14 ml/min/kg, 55% had a low-risk HFSS and 88% one-year event-free survival. Conversely, 22% of patients with a medium-risk HFSS had peak Vo2of > 14 ml/min/kg and demonstrated 85% one-year event-free survival. Therefore, intermediate-risk groups identified by either method include a sizeable portion of patients with an expected good prognosis on medical therapy alone. Identifying these patients can substantially impact the transplant list, considering that in our study, 78 (42%) of 187 of patients in the current era had peak Vo2between 10 and 14 ml/min/kg and 58 (31%) of 187 patients had a medium-risk HFSS.
Peak Vo2alone accurately predicted low- and high-risk patients, similar to a more comprehensive HFSS. However, unlike the favorable prognosis of intermediate-risk peak Vo2patients, the medium-risk HFSS patients, as a group, had 71% one-year event-free survival, identifying a group of patients who are likely to benefit from transplantation. However, HFSS in our study did not differentiate between medium- and high-risk patients. As mentioned earlier, a sizeable portion of patients with a medium-risk HFSS had preserved peak Vo2and, in turn, a better prognosis. When stratified by beta-blockers and defibrillator use, these patients had 90% one-year event-free survival when both therapies were used. Finally, beta-blockers therapy, in itself, may effect the calculation and, in turn, possibly the calibration of HFSS, as it impacts several of its constituent variables. This is demonstrated by the fact that despite a comparable ejection fraction, sodium concentration, New York Heart Association functional class, prevalence of intraventricular conduction delay, and peak exercise Vo2, the distribution of patients in the three HFSS groups was different in the current versus past era. Thus, for a similar degree of HF burden, patients are likely to have a different HFSS calculation now. All of these features suggest that HFSS in the current era needs to be modified to better stratify risk across the HF spectrum.
Overall, patients with peak Vo2of 10 to 14 ml/min/kg had survival rates that were comparable to those after transplantation; however, some patients in this group may have a worse prognosis. As shown in Table 2, however, the majority of patients with peak Vo2between 10 and 14 ml/min/kg actually have a low-risk HFSS and can safely be followed on medical therapy alone. One possible algorithm for selection of transplant candidates is suggested in Figure 4. Using this approach, one can possibly reduce the transplant list by approximately one-third without jeopardizing patient outcomes. Alternatively, newer prognostic schemes may be developed using different markers.
Our recommendations are based on one-year outcomes. It is possible that HF patients may have worse outcomes, as compared with transplantation, in the longer run (over 5 to 10 years), especially considering that the mortality rate after transplantation is highest in the first year after transplantation but decreases significantly thereafter. However, an alternate approach would be to continue medical therapy for patients who have a good short-term prognosis and, if and when their disease progresses to the point where short-term outcomes are likely to be worse than transplantation, then to proceed with evaluation and listing. Because medical urgency at the time of transplantation has only a mild impact on post-transplant mortality (first-year post-transplant survival 84.8 ± 0.7% vs. 87.4 ± 1.0% for status 1 vs. status 2), it may be feasible to pursue such an approach (1). Many of these patients are likely to remain stable for long periods of time, obviating the need for transplantation; some may never need it. Ongoing periodic reassessments should identify the group of patients who continue to deteriorate on medical therapy in order to evaluated and listed for a transplant in a timely manner.
There are several limitations to our study. First, the HF patients from the two eras were similar in terms of multiple clinical characteristics (e.g., age, gender, race, ejection fraction, serum sodium, intraventricular conduction delay, peak Vo2); therefore, we could compare outcomes between these groups. On the contrary, transplanted patients are a distinct group that may or may not be directly compared to the HF patients with respect to outcomes. Patients who were transplanted may represent a group that was sicker than the average HF patients, making their improved outcomes, compared with conventional HF treatment, even more impressive in the long run. On the other hand, the transplanted patients may represent a carefully selected group of HF patients with less comorbidity burden and, in turn, a better expected prognosis. Hence, in the absence of a prospectively designed trial, direct comparison between an average HF group of patients and those who get transplanted has inherent limitations. Other limitations include the fact that this is a single-center experience with retrospective data collection. Not all eligible patients were on beta-blockers at recommended doses in the current era. There were only four patients who had a biventricular pacemaker placed. The use of beta-blockers, defibrillators, and biventricular pacemakers is expected to increase in the future. It remains unknown how these trends would impact our results. Similarly, there may be differences in outcomes between different types of pacemakers, and we do not have the information on the overall use and type of pacemakers in our patients, except for those who received biventricular pacemakers (25). Finally, our study only addresses survival and does not assess quality of life for patients undergoing medical therapy versus transplantation.
Our study shows that the prognosis of HF patients has significantly improved in the current era of therapy. Many patients with intermediate risk now have survival comparable to that after transplantation. These trends are likely to continue to improve as newer therapies are discovered and current therapies are more widely utilized. These data suggest that the evaluation and listing criteria for heart transplantation in the current era need re-evaluation.
- heart failure
- heart failure survival score
- left ventricular assist device
- United Network for Organ Sharing
- oxygen consumption
- Received June 3, 2003.
- Revision received August 23, 2003.
- Accepted August 26, 2003.
- 2004 American College of Cardiology Foundation
- Mudge G.H.,
- Goldstein S.,
- Addonizio L.J.,
- et al.
- Costanzo M.R.,
- Augustine S.,
- Bourge R.,
- et al.
- Hunt S.A.,
- Baker D.W.,
- Chin M.H.,
- et al.
- Bristow M.R.,
- Gilbert E.M.,
- Abraham W.T.,
- et al.,
- the MOCHA Investigators
- Aaronson K.D.,
- Schwartz J.S.,
- Chen T.M.,
- et al.
- Mancini D.M.,
- Eisen H.,
- Kussmaul W.,
- et al.
- Aaronson K.D.,
- Mancini D.M.
- Zugck C.,
- Haunstetter A.,
- Kruger C.,
- et al.
- Stevenson L.W.,
- Sietsema K.,
- Tillisch J.H.,
- et al.
- Deng M.,
- De Meester J.M.J.,
- Smits J.M.A.,
- et al.