Author + information
- Received August 13, 1998
- Revision received November 10, 1998
- Accepted December 23, 1998
- Published online April 1, 1999.
- ↵*Reprint requests and correspondence: Keith Aaronson, University of Michigan Health System, Division of Cardiology, Taubman 3910, Ann Arbor, Michigan 48109-0366
The study aimed to determine the risk of death or urgent transplant for patients who survived an initial 6 months on the outpatient heart transplant waiting list when criteria emphasizing reduced peak oxygen consumption are used for transplant candidate selection.
Waiting time is a key criterion for heart donor allocation. A recent single-center investigation described decreasing survival benefit from transplant for patients who survived an initial 6 months on the outpatient waiting list.
Kaplan-Meier survival analyses were performed for 80 patients from the Hospital of the University of Pennsylvania (HUP) listed from July 1986 to January 1991, and 132 patients from Columbia-Presbyterian Medical Center (CPMC) listed from September 1993 to September 1995. Survival from the time of outpatient listing for the entire group (ALL) was compared to subsequent survival from 6 months onward for those patients who survived the initial 6 months after placement on the outpatient list (6M). Both urgent transplant and left ventricular assist device implantation were considered equivalent to death; elective transplant was censored.
Survival for 6M was not significantly better than ALL at HUP (subsequent 12 months: 60 ± 7 vs. 60 ± 6% [mean ± SD]; p = 0.89) nor at CPMC (subsequent 12 months: 60 ± 6 vs. 48 ± 5%; p = 0.35). Survival for 6M at both centers was substantially lower than survival following transplant from the outpatient list in the United States in 1995.
When high-risk patients are selected for nonurgent transplant listing, mortality remains high, even among those who survive the initial six months after listing. Time accrued on the waiting list remains an appropriate criterion for donor allocation.
The success of cardiac transplantation as a treatment for severe heart disease has resulted in a severe shortage of available donor organs (1,2). Suggested solutions to this problem have included methods of expanding the donor pool (3–11), various surgical alternatives to transplant (12–17), and more restrictive transplant candidate criteria (1,18). Once a patient is identified as a transplant candidate he or she is listed with the United Network for Organ Sharing (UNOS). The present heart donor allocation scheme awards points for waiting time, priority status (i.e., status 1 for left ventricular assist device patients or hospitalized patients on inotropic support in an intensive care unit, and status 2 for all others), weight, blood type and geography. Thus, the longer a patient remains on the waiting list, the more likely he or she is to receive the next donor organ. As heart failure is a progressive disease, the application of waiting time as a key factor in the distribution of organs is logical. However, a study by Stevenson et al. (19)implied that prioritizing on the basis of accumulated waiting time might not be appropriate for less urgent (UNOS status 2) candidates. In that study, outpatient transplant candidates who survived the initial six months on the waiting list without receiving a transplant experienced a survival rate in the subsequent year that was only marginally worse than would have been expected following cardiac transplantation. We reexamined this question by comparing the survival of UNOS 2 listed transplant candidates to the subsequent survival of those patients who survived six months on the waiting list at two transplant centers where strict criteria for acceptance of candidates were used. We also investigated the incidence and outcome of candidates who experienced substantial clinical improvement during the initial six months and were therefore deactivated from the waiting list.
Two hundred sixty-eight (268) ambulatory patients aged ≤70 were evaluated for transplant candidacy at the Hospital of the University of Pennsylvania (HUP) between July 1, 1986, and December 31, 1991. At Columbia-Presbyterian Medical Center (CPMC), 199 ambulatory patients ≤70 were evaluated between July 1, 1993, and July 1, 1995. Eighty (80) patients (69 men and 11 women), age 54 ± 9 years (mean ± SD) at HUP and 132 patients (103 men, 29 women) age 52 ± 11 years at CPMC were accepted for UNOS 2 transplant listing. Clinical characteristics and survival statistics for both samples have been described previously (20).
At both centers, patients accepted for transplant listing had New York Heart Association (NYHA) class III or class IV functional limitations and severely reduced left ventricular systolic function. A reduced peak exercise oxygen consumption (V̇o2) was used to guide the selection of potential transplant candidates. Peak V̇o2<14 ml·kg−1·min−1was considered a strong indication for transplant (21). Patients with higher peak V̇o2measurements were accepted as transplant candidates if the percent of predicted peak V̇o2was less than 40% (22,23)or for malignant ventricular arrhythmias uncontrolled with drugs or an implantable cardioverter-defibrillator, or unmanageable fluid overload requiring recurrent admissions for heart failure. Peak V̇o2was measured during maximal treadmill exercise testing using a modified Naughton protocol and a metabolic cart (Sensor-Medics, Anaheim, California).
Patients were followed prospectively in the heart failure clinics at HUP and CPMC at <3-month intervals. Measurements of peak V̇o2were repeated no later than six months after initial listing, or earlier if medical therapy was modified or subjective clinical improvement was noted. Patients for whom placement on the waiting list was based on a reduced peak V̇o2were removed from the waiting list if they experienced significant subjective improvement, which was confirmed by an increase in peak V̇o2of ≥2 ml·kg−1·min−1with achievement of an absolute peak V̇o2of ≥15 ml·kg−1·min−1on two consecutive studies performed at least two weeks apart. The date of transplant, left ventricular assist device (LVAD) implantation, or death without transplant was recorded for each patient, as was the UNOS priority status (1 or 2) at the time of transplant. For patients who remained alive and untransplanted, follow-up was discontinued on January 1, 1993, at HUP and on March 1, 1998, at CPMC.
All data are represented as mean ± SD. Comparisons of clinical variables between patients at HUP and CPMC were made by independent sample t, Wilcoxon rank-sum test, or the Fisher exact tests as appropriate. Disease progression resulting in critical clinical deterioration or death—defined as death without transplant, status 1 transplant (i.e., receiving mechanical or inotropic support prior to transplantation) or LVAD insertion—was the outcome event for this study. Follow-up was complete in all patients.
To investigate whether clinical characteristics present at listing could discriminate between patients who developed an outcome event (as defined above) during the following six months and those who did not, the clinical characteristics of these two groups were compared. Patients whose outcome was censored at <6 months (i.e., alive without a transplant or status 2 at the time of transplant) were not included in either group.
At each center, “survival” (i.e., freedom from an outcome event) from the time of listing for the entire group (ALL) was compared to subsequent survival from six months onward for the subgroup of patients (6M) who survived the initial six months after listing and were not removed from the waiting list for being “too well for transplant,” using the Kaplan-Meier method and the log-rank test. UNOS status 2 transplant was treated as a censored observation (i.e., removed from follow-up “alive” at the time of transplant). The power to detect a clinically significant difference in survival between ALL and 6M was calculated for the data from each center. An overall “experimentwise” power to detect a significant difference in either study was also determined using the addition rule of probability: power (HUP or CPMC) = power (HUP) + power (CPMC) − power (HUP and CPMC).
Power calculations were performed with PASS statistical software, version 6.0 (Number Cruncher Statistical Systems, Kaysville, Utah). All other calculations were performed with SAS statistical software, version 6.12 for Macintosh (SAS Institute, Cary, North Carolina). Statistical significance was defined as p ≤ 0.05, without adjustment for multiple comparisons.
Clinical characteristics of the study populations are shown in Table 1, and are consistent with the typical characteristics of UNOS status 2 patient samples: mainly men in their mid-50s with ischemic cardiomyopathy, NYHA class III heart failure and severely reduced LVEF and peak V̇o2. In comparison to the HUP cohort, the CPMC group had a larger proportion with nonischemic etiology and higher LVEF, but was similar with respect to age, gender, NYHA class, peak V̇o2, and the proportion of patients receiving an angiotensin-converting enzyme (ACE) inhibitor, digoxin and a diuretic.
There were no statistically significant differences in patient outcomes during the initial six months after placement on the UNOS status 2 list between the HUP and CPMC samples (Table 2). At six months, there were trends toward more patients remaining alive and untransplanted (68% vs. 55%, p = 0.107) and fewer patients dying or receiving a UNOS 1 transplant or LVAD (22% vs. 32%, p = 0.061) at HUP versus CPMC.
Only 4% of patients at each center (three patients at HUP and five patients at CPMC) were removed from the waiting list because of clinical and objective improvement during the initial six-month period. None of the delisted patients at HUP experienced an outcome event during follow-up periods of 491, 624 and 933 days. Two of the five delisted patients at CPMC survived without an outcome event during follow-up periods of 857 and 927 days. The remaining three CPMC patients died suddenly after 214, 824 and 846 days of follow-up.
There were 51 six-month survivors at HUP and 68 at CPMC. Outcomes for the 6M groups are shown in Table 3. Survival for the 6M group was not significantly different from overall survival for either sample (Figs. 1 and 2). ⇓⇓At HUP, survival for 6M versus ALL was 76 ± 6% versus 77 ± 5% at 6 months, 60 ± 7% versus 60 ± 6% at 12 months, and 39 ± 9% versus 33 ± 7% at 24 months, respectively (p = 0.89). Corresponding survival rates at CPMC were 75 ± 6% versus 62 ± 4% at 6 months, 60 ± 8% versus 48 ± 5% at 12 months, and 35 ± 5% versus 32 ± 7% at 24 months, respectively (6M vs. ALL, p = 0.35).
If the 12% absolute improvement in survival (60% vs. 48%) at one year at CPMC is considered clinically meaningful, our study had 60% power to have found this difference to be statistically significant at the 0.05 level. The power to detect the same improvement in survival in the HUP dataset was only 45%, due to the smaller sample size. However, because these two datasets are independent, the power to have found this 12% difference in either sample was 78% (0.6 + 0.45 − (0.6)(0.45)).
Superimposed on each figure are survival data from the UNOS Scientific Registry for all adult patients who were UNOS status 2 at the time of transplantation in 1995. The 88 ± 1% and 86 ± 1% actuarial survival rates at six and 12 months after UNOS status 2 transplant were significantly better than the survival rates for 6M at both HUP and CPMC (both p < 0.0001).
Clinical characteristics of 6-month survivors (including patients who were delisted during the first 6 months) and nonsurvivors are shown in Table 4. No statistically significant differences were present for individual clinical characteristics. Although there was a trend toward higher LVEFs in survivors at HUP (18 ± 6% vs. 15 ± 6%, p = 0.055), no difference was present for the CPMC group (20 ± 8 [survivors] vs. 21 ± 10 [nonsurvivors], p = 0.55).
The study suggests that when selection criteria emphasizing reduced peak V̇o2are used to select outpatient candidates for cardiac transplantation, a high risk of death or urgent transplantation is present well beyond the initial six months after placement on the waiting list. The UNOS status 2 patients who survive the initial six months of transplant listing without receiving a donor organ continue to experience a high risk of death or need for urgent transplant. Because this risk remains well in excess of the risk of death following cardiac transplantation, the present policy of UNOS 2 listing prioritization, which awards credit for time spent on the waiting list, continues to be appropriate.
Over the past decade there has been a consistent rise in the referral of potential heart transplant candidates. This has resulted in the growth of the transplant waiting list from 1,030 in 1988 to 3,698 in 1996 (2). During this same period, the proportion of patients waiting more than one year has increased from 12.5% to 50.6% (2). The supply of donor organs has not kept pace—in fact, heart donation has plateaued at approximately 2,500 annually for each of the past four years (2). Various approaches to addressing this supply-demand imbalance have been suggested: lowering the age cutoff for transplant candidates (1), accepting hearts from older donors (3–7), accepting donor hearts with preexisting coronary artery disease for combined coronary artery bypass grafting and transplant (8,9), cardiomyoplasty (12–14), mitral valve repair alone (17), or with partial left ventriculectomy (16), permanent LVAD implantation (15,24)and xenotransplantation (10).
Another approach would involve changing the current U.S. organ allocation system for outpatient candidates. At present, excluding consideration of blood group and weight range, outpatient prioritization is awarded solely on the basis of waiting time. This is rational and equitable if the mortality rate remains high among patients who have remained on the waiting list without receiving a transplant. Referral to a transplant center and subsequent placement on the waiting list often follows a recent deterioration, so mortality during the following few months is expected to be high. However, careful medical management at a transplant center might result in substantially improved survival. Studying this question, Stevenson and colleagues at UCLA observed a 67% one-year actuarial survival without transplantation for 214 outpatients from the time of placement on the waiting list. However, for outpatients who had already survived 6 months from the time of listing, survival without transplant was 83% over the subsequent year (19). This was only slightly worse than the 88% one-year survival following cardiac transplantation at their institution, and compares favorably to the 83% one-year survival after heart transplantation in the United States (2).
A study by Kao et al. (25)seemed to confirm the UCLA groups findings in that subsequent mortality among patients who survived an initial six months on an outpatient waiting list was only 6% at six months and 12% at 12 months. However, when patients on inotropic support at the time of transplantation were considered to have died (while still awaiting transplant) three months after the date of their transplant, the overall pretransplant mortality rate increased to 13% at six months and 35% after 12 months.
Neither of the aforementioned studies used rigorous selection criteria emphasizing reduced peak V̇o2. When these criteria are applied this phenomena of altered short- versus long-term survival may not persist. Congestive heart failure (CHF) is a progressive disease. If patients are placed on the waiting list only after medical therapy has been optimized, the mortality rate on the waiting list should continue to be substantial as the mean waiting time increases. Moreover, clinical improvement of a sufficient degree to allow candidate deactivation should be infrequent. In our series it is not surprising that subsequent one-year survival for six-month survivors at HUP and CPMC (60 and 62%, respectively) was similar to survival for each entire cohort.
The survival time-censoring methodology used by Stevenson and colleagues (19)may have given a false impression of a good outcome for 6-month survivors in the UCLA study. In both our studies and the one at UCLA survival time was censored at the time of UNOS 2 transplant. However, 41% of the UCLA sample (and 49% of those actually listed for transplant) received a UNOS 2 transplant versus 29% at HUP and 23% at CPMC. When survival data on a patient are censored, actuarial survival methods assume that the patient’s probability of survival beyond the time of censoring would have been the same as that for patients for whom survival data can continue to be collected. While there is no a priori reason to believe that this was not true for the UCLA series, the large proportion of censored outcomes makes the analysis dependent on the survival experience of a relatively small proportion of patients.
Our study also differed from the study of Stevenson et al. (19)in that we treated UNOS status 1 and LVAD insertion as equivalent to death, whereas UNOS 1 transplant was censored in the UCLA study (LVADs were not used in the UCLA study). This difference is reflected in lower event-free survival rates for the entire HUP and CPMC cohorts than for the UCLA cohort. However, censoring for UNOS status 1 transplant and LVAD insertion is very misleading. Patients who require intravenous inotropes or an LVAD to survive cannot be expected to have the same likelihood of survival without a transplant as patients who are less ill and remain on the outpatient waiting list. Although treating UNOS 1 transplant or LVAD as equivalent to death is also problematic—patients receiving these interventions have not died—the expected survival experience of these patients is much closer to that of patients who have died than it is to patients who remain on the UNOS status 2 waiting list.
It is also important to consider whether the clinical approach to care taken by Stevenson et al. (26–29)was responsible for the improved outcomes for six-month survivors. These investigators have suggested that “tailored therapy,” a program of high-dose diuretics and vasodilators titrated to hemodynamic goals, results in improved outcomes when applied to patients with advanced heart failure referred for cardiac transplant evaluation (26–28). However, it is likely that many of these patients were receiving suboptimal therapy when initially referred for evaluation, and it is not clear whether any benefit derived from tailored therapy would not have followed conventional optimization of medical therapy with digoxin, a diuretic and target doses of an ACE inhibitor. This remains only speculative as no randomized clinical trial has ever compared tailored therapy to contemporary standard heart failure therapy.
No patient received an urgent (UNOS status 1) transplant, and nearly all of the deaths were sudden for patients who survived the initial six months after listing in the UCLA series. In our two samples, the majority of events were UNOS status 1 transplants in patients who experienced hemodynamic deterioration requiring inotropic bridging to transplant. This is consistent with the advanced state of heart failure for listed patients in our samples and is also consistent with the experience reported by Kao et al. (26).
Both Stevenson et al. (29)and Levine et al. (30)have reported the frequent occurrence of listed patients who experience substantial objective and subjective clinical improvement allowing successful removal from the waiting list. However, clinical improvement warranting removal from the transplant waiting list was unusual in both of our samples. When objective criteria emphasizing peak V̇o2obtained after optimizing medical therapy are used to select ambulatory patients for placement on a transplant waiting list, substantial clinical improvement should be uncommon. Patients with a recent onset of a nonischemic cardiomyopathy constitute an important exception. At both centers, we frequently deferred transplant listing for a few months in these patients to allow for spontaneous improvement. Nevertheless, we believe that all listed patients should undergo serial reevaluations at 6-month intervals (or sooner if there appears to be substantial improvement) to identify patients who may be removed from the waiting list.
The use of UNOS status 1 transplant as an end point equivalent to death represents a limitation of the study. However, the need for urgent transplantation or implantation of an LVAD does signify critical clinical deterioration and disease progression. Whereas combining UNOS 1 transplant and death overstates the “true” mortality rate that would exist if heart transplant were not available, using death alone as the outcome would understate the “true” mortality. This study is also limited in that the data collected reflects the practice of two large transplant centers and not the entire country.
Clinical implications and conclusions
From 1988 to 1995 median waiting time for a donor heart increased from 122 days to 213 days (2). In a multivariable analysis of factors predictive of waiting time in the UNOS registry from April 1994 through April 1996, patients who spent all of their waiting time as outpatients had a 7.3-fold longer wait (J.M. Chen, personal communication, 1997). Patients experiencing these prolonged waiting times remain at substantially increased risk of dying and should continue to receive priority for a donor organ.
We have shown that high-risk patients can be identified when objective criteria emphasizing reduced peak V̇o2are used to select transplant candidates. For such patients, mortality remains high, even among those who survive an initial six months after placement on the waiting list. Improvement of a sufficient degree to allow removal from the transplant list is infrequent in these circumstances. Therefore, time accrued on the outpatient waiting list remains an appropriate criterion for donor allocation.
☆ Dr. Aaronson was supported by a Clinical Investigator Development Award (K08-HL02829) from the National Heart, Lung and Blood Institute, National Institutes of Health.
Portions of this work were presented in abstract form at the Fourteenth Annual Meeting and Scientific Session of the International Society for Heart and Lung Transplantation, March 1996.
- entire cohort at HUP (or CPMC)
- angiotensin-converting enzyme
- congestive heart failure
- Columbia-Presbyterian Medical Center
- Hospital of the University of Pennsylvania
- left ventricular assist device
- New York Heart Association
- Peak V̇o2
- peak exercise oxygen consumption
- United Network for Organ Sharing
- patients who survived for 6 months on the UNOS status 2 waiting list without receiving UNOS status 1 transplant or LVAD
- Received August 13, 1998.
- Revision received November 10, 1998.
- Accepted December 23, 1998.
- American College of Cardiology
- Stevenson L.W,
- Warner S.L,
- Steimle A.E,
- et al.
- ↵UNOS. 1997 Annual Report of the US Scientific Registry for Transplant Recipients and the Organ Procurement and Transplantation Network. Data 1988–1997: UNOS, Richmond, VA, and the Division of Transplantation, Bureau of Health Resources Development, Health Resources and Service Administration, US Department of Health and Human Services, Rockville, MD, 1998.
- Michler R.E,
- Chen J.M,
- Mancini D.M,
- Reemtsma K,
- Rose E.A
- Oz M.C,
- Pinsky D.J,
- Koga S,
- et al.
- Levin H.R,
- Weisfeldt M.L
- McCarthy J.F,
- McCarthy P.M,
- Starling R.C,
- et al.
- ↵Bolling SF, Pagani FD, Deeb GM, Bach DS. Intermediate-term outcome of mitral reconstruction in cardiomyopathy. J Thorac Cardiovasc Surg 1988;115:381–6; discussion 387–8.
- Stevenson L.W,
- Warner S.L,
- Hamilton M.A,
- et al.
- Stevenson L.W,
- Hamilton M.A,
- Tillisch J.H
- Aaronson K.D,
- Schwartz J.S,
- Chen T.M,
- Wong K.L,
- Goin J.E,
- Mancini D.M
- Mancini D.M,
- Eisen H,
- Kussmaul W,
- Mull R,
- Edmunds L.H,
- Wilson J.R
- Aaronson K.D,
- Mancini D.M
- Stelken A.M,
- Younis L.T,
- Jennison S.H,
- et al.
- Kao W,
- McGee D,
- Liao Y,
- et al.
- Stevenson L.W,
- Steimle A.E,
- Fonarow G,
- et al.