HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Author home page(s): Jonathan M. Chen Eric A. Rose Robert E. Michler Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Chen, J. M. Articles by Michler, R. E. Search for Related Content PubMed PubMed Citation Articles by Chen, J. M. Articles by Michler, R. E.

Ann Thorac Surg 1996;61:570-575
© 1996 The Society of Thoracic Surgeons

---

Original Article: Cardiovascular

Multivariate Analysis of Factors Affecting Waiting Time to Heart Transplantation

Departments of Surgery and Medicine, Columbia-Presbyterian Medical Center, New York, New York

Accepted for publication September 16, 1995.

	Abstract

Top Footnotes Abstract Introduction Material and Methods Results Comment References

 
Background. The growing clinical success of cardiac transplantation has resulted in a dramatic increase in the number of patients referred and subsequently listed for cardiac transplantation. Paradoxically, in the presence of a limited donor organ pool, such expansion has increased both the waiting time for transplantation and the number of patients dying while on the waiting list.

Methods. We performed univariate and multivariate analyses of the waiting times of 301 patients listed for transplantation using a Cox proportional hazards model to evaluate the simultaneous effect of multiple variables on the waiting time of heart transplant candidates. Variables considered included age, sex, race, blood type, weight at listing, United Network for Organ Sharing (UNOS) status at listing, UNOS status at transplantation, and proportion of time on the waiting list as UNOS status 1.

Results. The mean waiting time for patients ultimately having transplantation was 170.2 ± 206.0 days; the median waiting time was 103.5 days. Age, sex, weight, blood type, and percent of time as UNOS status 1 all had a significant impact on waiting time in the univariate analysis. By multivariate analysis, proportion of time as UNOS status 1, lower weight at listing, and blood type AB were all highly associated as predictors of a shorter waiting time. Weight at listing represented a continuous variable whose risk ratio for a shorter waiting time correlated in such a way that the risk of a longer waiting time increased 2.3 per 22.5-kg (50-pound) increase in weight. Blood types A and B, although associated with a shorter waiting time, correlated less strongly than the other three variables.

Conclusions. Our findings from this multivariate analysis demonstrate that UNOS status, blood type, and weight were the variables that most strongly affected overall waiting time for transplantation. It is our hope to define more accurately a group of patients with both a high likelihood of a long waiting time and a prohibitive risk of death while on the waiting list, who therefore may benefit from surgical alternatives to transplantation.

	Introduction

Top Footnotes Abstract Introduction Material and Methods Results Comment References

 
The growing success of clinical transplantation has dramatically increased the number of patients referred and subsequently listed for cardiac transplantation. However, the limited donor pool currently available has created an ever-growing demand for donor organs. Over the past 10 years, the number of patients on the waiting list for heart transplantation, the average waiting time to transplantation, and thus the number of patients dying before a suitable donor organ can be found have progressively increased. In 1993 alone, 3,776 patients were listed by the United Network for Organ Sharing (UNOS) for heart transplantation, 1,028 (27.2%) of whom were still waiting for organs at year's end; the median waiting time was 207 ± 8.8 days. These values do not compare favorably with data from 5 years earlier in which only 2,764 patients were listed for transplantation with a median waiting time of 108 days (Fig 1) (UNOS, Richmond, VA; personal communication).

View larger version (32K): [in this window] [in a new window]   Fig 1. . Increase in both number of patients listed for heart transplantation annually with United Network for Organ Sharing and median waiting time for the years 1988 through 1993.

In a previous univariate analysis, we [2] evaluated variables that had an impact on a potential recipient's waiting time to heart transplantation. We undertook the current study to evaluate the simultaneous effect of multiple variables on the waiting time to heart transplantation. It is our hope that more accurate characterization of patients at highest risk for longer waiting times may help to reduce mortality of patients on the waiting list and also may allow early application of therapeutic alternatives.

	Material and Methods

Top Footnotes Abstract Introduction Material and Methods Results Comment References

 
Patient Selection
From September 17, 1991, through September 28, 1994, 301 patients were listed for cardiac transplantation at the Columbia-Presbyterian Medical Center. All patients were registered with the national database of UNOS. No patient who was listed for transplantation during this period was excluded from the initial analysis.

Study Variables
First a univariate analysis was performed to assess the effects of individual variables on the waiting times of heart transplant candidates. Variables considered included age, sex, race, blood type, weight at listing, UNOS status at listing, UNOS status at transplantation, and proportion of time on the waiting list as UNOS status 1. The proportional hazards model of Cox [3] was applied to this cohort in an effort to assess the simultaneous effect of multiple variables on the waiting times of heart transplant candidates. The variables considered were the same as for the univariate analysis.

The outcome variable of total waiting time was calculated as the total time listed as active (either status 1 or status 2) on the UNOS database. Time accrued while listed as a status 7 (inactive) was not considered for this analysis. End-point variables that resulted in censorship from the model included death and removal from the list. Proportion of time on the waiting list as a UNOS status 1 patient was calculated by dividing the total amount of time listed as UNOS status 1 by the total amount of time actively listed for transplantation.

Statistical Methods
Univariate analysis was performed using the log rank ² test for homogeneity across the strata. In this univariate analysis, continuous variables were subdivided into groups for the purpose of comparative evaluation. Multivariate analysis was performed using the Cox proportional hazards model. Outcome and end-point variables for both studies were employed as already described. Continuous variables were not stratified in the multivariate analysis, and significance was defined as a p value of less than 0.01.

	Results

Top Footnotes Abstract Introduction Material and Methods Results Comment References

 
Patient Population
Of the 301 adults listed for transplantation between Septebmer 17, 1991, and September 28, 1994, 142 underwent heart transplantation. At the close of the study interval, 33 patients (11.0%) were still actively waiting, 60 (19.9%) had died, 53 (17.6%) were listed as status 7 (inactive), and 13 (4.3%) had been permanently removed from the waiting list.

In the total cohort, 68 patients (22.6%) were women, and 233 (77.4%) were men. Blood type O was the most common (122 patients, 40.5%), followed by blood type A (116 patients, 38.5%), blood type B (44 patients, 14.6%), and blood type AB (19 patients, 6.3%). There were 244 white patients (81.1%) and 57 (18.9%), nonwhite patients. Age ranged from 20.5 years to 75.7 years with a mean age of 56.7 ± 12.9 years. Weight ranged from 42.57 kg to 133.83 kg with a mean weight of 73.94 ± 16.07 kg.

The patients listed spent, on average, 24.2% of total waiting list time as UNOS status 1. Of those patients who ultimately underwent transplantation, the mean time of active listing was 170.5 ± 205.9 days (range, 0 to 1,165 days), and the median waiting time was 103.5 days. In comparison, the median waiting time for the 301 patients in the total cohort was 287 days.

Of the 60 patients who died while on the waiting list, 50 (83.3%) were male, and 10 (16.7%) were female. Twenty-nine (48.3%) of these patients were blood type O, 23 (38.3%) were blood type A, 7 (11.7%) were blood type B, and 1 patient (1.7%) was blood type AB. Forty-nine (81.7%) were white and 11 (18.3%), nonwhite.

Waiting Time Analysis
Results of the univariate analysis are summarized in Figures 2 through 6. As shown percent of time as UNOS status 1, blood type, age, sex, and weight all had a significant impact on waiting time.

View larger version (15K): [in this window] [in a new window]   Fig 2. . Kaplan-Meier plot representing the percentage of patients waiting by the variable proportion of time on waiting list as United Network for Organ Sharing (UNOS) status 1.

View larger version (20K): [in this window] [in a new window]   Fig 3. . Kaplan-Meier plot representing the percentage of patients waiting by the variable blood type.

View larger version (18K): [in this window] [in a new window]   Fig 4. . Kaplan-Meier plot representing the percentage of patients waiting by the variable age.

View larger version (15K): [in this window] [in a new window]   Fig 5. . Kaplan-Meier plot representing the percentage of patients waiting by the variable sex.

View larger version (19K): [in this window] [in a new window]   Fig 6. . Kaplan-Meier plot representing the percentage of patients waiting by the variable weight.

where h(t'X) = waiting time hazard function; h₀(t) = baseline hazard function; A, B, and AB = blood types; WT = weight; and UNOS = proportion of time spent as UNOS status 1.

Results of the multivariate analysis are presented in Table 1. In summary, the proportion of time as UNOS status 1, a lower weight at listing, and blood type AB all were highly associated as predictors of a shorter waiting time as evidenced by the p value of 0.0001 and the risk ratio. United Network for Organ Sharing status was the variable most highly associated with a shorter waiting time. Weight at listing represented a continuous variable whose waiting time risk ratio for a shorter waiting time correlated in such a way that the relative risk increased 2.3 per 22.5-kg (50-pounds) decrement in weight. Thus, for every increase in weight of 22.5 kg, an individual's likelihood of having a longer waiting time was increased approximately twofold. Blood types A and B, although associated with a shorter waiting time, correlated less strongly than the three variables, proportion of time as UNOS status 1, lower weight at listing, and blood type AB.

	Comment

Top Footnotes Abstract Introduction Material and Methods Results Comment References

 
Despite expansion of the donor pool to include older donors as well as more marginal donors, the growing clinical success of cardiac transplantation paradoxically has resulted in an increase in the waiting time to transplantation, with a subsequent increase in mortality of patients on the waiting list. Thus, considerable debate has arisen regarding the current equity of organ distribution, the appropriateness of transplantation versus medical management in select instances, and the ideal window for intervention with ``bridging'' alternatives. Although previous univariate analyses have identified factors that potentially contribute to longer waiting times, we undertook the current multivariate analysis to assess the simultaneous effect of multiple variables on the waiting times of heart transplant candidates at a single institution that performs approximately 80 transplants a year [2].

Similar, more extensive investigations regarding waiting time analysis have previously been undertaken with large cohorts of renal transplant recipients [4]. However, unlike renal transplantation where hemodialysis allows time for the matching of more subtle immunologic factors prior to transplantation, in cardiac transplantation, the lack of a similar, widely available alternative has rendered weight, blood type, and clinical urgency the primary criteria on which organ distribution is based.

In a previous study, we [2] evaluated similar variables in a univariate analysis and demonstrated a higher probability of shorter waiting times for women, adults weighing less than 81 kg, and patients with blood types AB, A, and B. These findings were confirmed in the univariate analysis performed here using a larger cohort of potential recipients. In addition, significant differences in waiting times were demonstrated with our univariate analysis in regard to age and proportion of time as UNOS status 1. The value of a multivariate analysis, however, was revealed by our current finding that no waiting time advantage was rendered on the basis of sex or age. This result suggests that our earlier univariate analysis may have reflected, for example, sex differences in overall weight range rather than a true sex bias in waiting time to transplantation.

For those two continuous variables analyzed in the multivariate model (proportional UNOS status waiting time and weight at listing), comparative waiting time risk ratios could be generated by maintaining all other variables constant. First, the proportion of time spent as UNOS status 1 was the variable most highly associated with waiting time (risk ratio = 41.7). Thus, given two individuals of equivalent weight and blood type, patient 1 who has spent 25% and patient 2 who has spent 75% of total waiting list time as UNOS status 1, the latter would have a 6.5-fold greater likelihood of receiving a heart than the former (h(t'X)₂/h(t'X)₁ = exp 1.87 = 6.5). Second, in regard to weight at listing, the multivariate analysis demonstrated an approximate twofold increase in waiting time for every 22.5 kg increase in weight.

The current UNOS policy confers organs first on the basis of clinical urgency and then, for patients within each weight and blood type category, on the basis of waiting time accrued. Patients listed as UNOS status 1 represent those candidates requiring more urgent transplantation (patients who are in an intensive care unit, who require inotropic support to maintain an adequate cardiac output, or who require cardiac or pulmonary mechanical assistance), whereas UNOS status 2 patients represent candidates who do not meet these criteria.

The original UNOS policy for organ distribution did not distinguish between total time accrued as a status 1 or status 2 candidate; rather it calculated waiting time as the cumulative time accrued from the time of initial listing [5]. In June 1993, however, it was proposed that patients accrue status 1 waiting time only while listed as status 1 and that patients accrue status 2 waiting time cumulatively from the initial time of listing [6, 7]. This measure attempted to have an impact on organ distribution equity by differentiating between (1) patients who have been chronically supported as outpatients but who acutely, albeit transiently, decompensate (and who may thus later improve without transplantation) and (2) patients who have accrued less time on the list but have required intensive care management throughout their hospital course. This policy was adopted by UNOS in February 1994 [8].

In an attempt to study most accurately the waiting times of patients in the scenario just described, we evaluated three variables to assess the effects of UNOS status on overall waiting time (status at listing, status at transplantation, and overall proportion of time as UNOS status 1). Both our univariate and multivariate analyses demonstrated (as already described) that the overall proportion of time a patient spent on the list as status 1 was the one variable of these three that was most highly associated with a shorter waiting time, a finding in support of the equity of the recent UNOS policy change.

A Markov chain model analysis by Stevenson and associates [10], however, has previously shown that the overall survival for all candidates is favored by the current system unless the perioperative mortality of candidates in critical condition were to exceed that of outpatients by more than fourfold. The conclusions from these previous modeling studies have emphasized the need to identify those patients at highest risk for sudden death while on the waiting list so that the sensitivity of the urgent and nonurgent status could be refined further, thus improving the efficacy of organ distribution.

Our findings not only support this notion but also complement its end point by attempting to identify better those patients who are at highest risk for longer waiting times. Indeed, recent studies by Saxton and colleagues [11] have demonstrated a decreasing benefit from transplantation for those patients whose survival on the waiting list exceeds 6 to 9 months. Although the current study is limited by the relatively small number of patients available, it is our hope that future similar multiinstitutional analyses including larger cohorts of patients and using multiple hazard analyses to describe waiting profiles more accurately may help to identify not only patients at highest risk for longer waiting times but also the changing risk of death (and thus the benefit of transplantation) for patients on the waiting list as waiting time increases.

The utility of such a profile would be twofold. As suggested by Stevenson and associates, there may be a small subset of patients for whom survival beyond 6 months on the waiting list is likely with aggressive medical management; for this small subset, overall survival may not be improved by transplantation. Thus, for those patients who in addition have a higher likelihood of a longer waiting time, long-term aggressive medical management may be the therapy of choice. Conversely, for the subset of patients who do not have a strong likelihood of surviving 6 months on the waiting list but for whom a longer waiting time is probable, the potential application of surgical alternatives to transplantation may be considered, including high-risk coronary revascularization, implantation of a bridging device (eg, a left ventricular assist device or xenograft), or application of a permanent alternative to transplantation. For this cohort of patients, the determination of waiting time estimates would aid in better defining a ``therapeutic window'' during which such alternative therapies might be most effective.

Major limitations of this study include the overall size of the cohort, the center-specific nature of the data, and most importantly, the assumption of the Cox proportional hazards model that the covariates studied are independent of time. Our goal in future studies ultimately will be to analyze the time-dependent change in the multivariate model so that outcome variables such as the risk of death on the waiting list may be evaluated for variance over time. In addition, we hope that the analysis of a larger cohort of potential recipients may help to derive a patient-specific algorithm to aid the clinician in estimating predicted waiting time and the relative risk of both transplantation and death. Insights into these and other transplantation outcome variables will aid in refining the basis on which organs are allocated, thus improving the overall survival of patients with end-stage heart disease.

	Footnotes

Top Footnotes Abstract Introduction Material and Methods Results Comment References

 
Please address all communications to the American Board of Thoracic Surgery, One Rotary Center, Suite 803, Evanston, IL 60201.

Address reprint requests to Dr Michler, Division of Cardiothoracic Surgery, Department of Surgery, Columbia-Presbyterian Medical Center, 177 Fort Washington Ave, New York, NY 10032.

	References

Top Footnotes Abstract Introduction Material and Methods Results Comment References

 

1. Stevenson LW, Hamilton MA, Tillisch J, et al. Decreasing survival benefit from cardiac transplantation for outpatients as the waiting list lengthens. J Am Coll Cardiol 1991;18: 919–25.[Abstract]
2. Michler RE, Chen JM, Mancini DM, Reemtsma K, Rose EA. Sixteen years of cardiac transplantation: the Columbia-Presbyterian Medical Center experience 1977–1993. In: Terasaki PI, Cecka JM, eds. Clinical transplants 1993. Los Angeles: UCLA Tissue Typing Laboratory, 1993:109–18.
3. Cox DR. Regression models and life tables. J R Stat Soc [B] 1972;34:87.
4. Sanfilippo FP, Vaughan WK, Peters TG, et al. Factors affecting the waiting time of cadaveric kidney transplant candidates in the United States. JAMA 1992;267:247–52.[Abstract]
5. Report of the Heart Transplantation Committee to the board of directors. Richmond, VA: The United Network for Organ Sharing. March 21, 1988:2–3.
6. Report of the Thoracic Organ Transplantation Committee to the board of directors. Richmond, VA: The United Network for Organ Sharing. March 3–4, 1993:2–3.
7. Report of the Thoracic Organ Transplantation Committee to the board of directors. Richmond, VA: The United Network for Organ Sharing. June 30–July 1, 1993:4.
8. Report of the Thoracic Organ Transplantation Committee to the board of directors. Richmond, VA: The United Network for Organ Sharing. Nov 3–4, 1993:1.
9. Stevenson LW, Warner SL, Steimle AE, et al. The impending crisis awaiting cardiac transplantation. Circulation 1994;89:450–7.[Abstract/Free Full Text]
10. Stevenson LW, Warner SL, Hamilton MA, et al. Modeling distribution of donor hearts to maximize early candidate survival. Circulation 1992;86(Supp 2):224–30.
11. Saxton LA, Stevenson WG, Middlekauff HR, et al. Predicting death from progressive heart failure secondary to ischemic or idiopathic dilated cardiomyopathy. Am J Cardiol 1993;72:62–5.[Medline]

This article has been cited by other articles:

				K. Lietz and L. W. Miller Improved Survival of Patients With End-Stage Heart Failure Listed for Heart Transplantation: Analysis of Organ Procurement and Transplantation Network/U.S. United Network of Organ Sharing Data, 1990 to 2005 J. Am. Coll. Cardiol., September 25, 2007; 50(13): 1282 - 1290. [Abstract] [Full Text] [PDF]

				C. C. Canver and J. Chanda Milrinone for long-term pharmacologic support of the status 1 heart transplant candidates Ann. Thorac. Surg., June 1, 2000; 69(6): 1823 - 1826. [Abstract] [Full Text] [PDF]

				D. G. Pennington, T. E. Oaks, and D. P. Lohmann Permanent ventricular assist device support versus cardiac transplantation Ann. Thorac. Surg., August 1, 1999; 68(2): 729 - 733. [Abstract] [Full Text] [PDF]

				D. J. Goldstein, M. C. Oz, and E. A. Rose Implantable Left Ventricular Assist Devices N. Engl. J. Med., November 19, 1998; 339(21): 1522 - 1533. [Full Text] [PDF]

This Article Abstract Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Author home page(s): Jonathan M. Chen Eric A. Rose Robert E. Michler Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Chen, J. M. Articles by Michler, R. E. Search for Related Content PubMed PubMed Citation Articles by Chen, J. M. Articles by Michler, R. E.