The Elderly Liver Transplant Recipient: A Call for Caution

Abstract

Objective

To determine whether liver transplantation is judicious in recipients older than 60 years of age.

Summary Background Data

The prevailing opinion among the transplant community remains that elderly recipients of liver allografts fare as well as their younger counterparts, but our results have in some cases been disappointing. This study was undertaken to review the results of liver transplants in the elderly in a large single-center setting. A secondary goal was to define, if possible, factors that could help the clinician in the prudent allocation of the donor liver.

Methods

A retrospective review of a prospectively maintained single-institution database of 1,446 consecutive liver transplant recipients was conducted. The 241 elderly patients (older than 60 years) were compared with their younger counterparts by preoperative laboratory values, illness severity, nutritional status, and donor age. Survival data were stratified and logistic regression analyses were conducted.

Results

Elderly patients with better-preserved hepatic synthetic function or with lower pretransplant serum bilirubin levels fared as well as younger patients. Elderly patients who had poor hepatic synthetic function or high bilirubin levels or who were admitted to the hospital had much lower survival rates than the sicker younger patients or the less-ill older patients. Recipient age 60 years or older, pretransplant hospital admission, and high bilirubin level were independent risk factors for poorer outcome.

Conclusions

Low-risk elderly patients fare as well as younger patients after liver transplantation. However, unless results can be improved, high-risk patients older than 60 years should probably not undergo liver transplantation.

The increasing shortage of donor livers for transplantation has led to long waiting times, pretransplant (waiting list) death, and heated, bitter debates on the most suitable way to allocate these organs. Median waiting times nationwide are approaching 200 days; in certain parts of the country, patients may wait up to 2 years for a liver transplant. ¹ These delays from waiting list placement to time of transplant inevitably translate into deaths of intended recipients. At our institution, waiting list deaths accrue at a rate of 20% of our adult liver transplant volume. Nationally, 10% of all adults die after being placed on the waiting list but before reaching transplant. ¹

One consequence of this shortage has been mounting acrimony in the transplant community about organ allocation, with the federal government recently issuing directives. ² Further, the governing board of the United Network for Organ Sharing accepted and implemented standardized criteria for placement on the liver transplant waiting list, with member-composed regional review boards to police the process. ¹ For the clinician, other consequences include carefully reexamining the results of liver transplantation and avoiding organ uses that produce suboptimal results. ^3,4

At our institution, we have become concerned that the results of liver transplantation in elderly recipients have in some cases been disappointing. In contrast, the transplant literature has steadfastly supported liver transplantation in elderly recipients. ^5–8 The current study was undertaken to review the results of liver transplants in the elderly in a large single-center setting. A secondary goal was to define, if possible, factors that could help the clinician in the prudent allocation of the donor liver.

METHODS

A retrospective review of a prospectively maintained computerized database was conducted. In a 13-year period, 1,446 consecutive liver transplants were performed at our institution. In this study, 1,205 recipients (83%) were younger than 60 years of age and 241 (17%) were 60 years or older. Initially, the following variables were compared between those two groups: preoperative status (home, hospital, intensive care unit), pretransplant levels of prothrombin time, albumin, and bilirubin, Child-Pugh score, nutritional status, and donor age. Patient survival data with each of these variables were calculated. In addition, logistic regression analyses were conducted in an attempt to isolate the influence of each of these variables on patient outcome. Finally, some of the same comparisons were repeated using different age groups (i.e., older than or younger than 50 years or 65 years) in an attempt to define the age, if any, at which the liver transplant recipient moves into a higher-risk group.

These variables were chosen for study because of known outcome relationships. For example, pretransplant hospital admission correlates highly with a poorer outcome, as does poor nutritional status and, in selected cases, advanced donor age. ^9–11 The laboratory values used were identified as other surrogates for the degree of the patient’s pretransplant decompensation, as was the Child-Pugh score. In addition, these laboratory values (as key components of the Child-Pugh score) and the Child-Turcotte-Pugh score itself warrant close scrutiny because they have such an important impact on the new United Network for Organ Sharing listing criteria. ¹ Historically, though, the ability to predict posttransplant outcome based on pretransplant factors has been poor. ¹²

RESULTS

We reviewed 1,446 consecutive patients: 1241 (86%) were 60 years or younger at the time of transplant, 137 (9%) were 61 to 65 years old, and 68 (5%) were 66 years or older. The vast majority of patients (91.7%) were in their fourth decade of life or older (median age 48 years, mean 47.2 years). Twenty (1.4%) were pediatric patients (17 years or younger). Men and women were equally represented. Only 10% of these patients went on to undergo retransplantation. Data on patient demographics are shown in Table 1. Patient follow-up was 46 months (median) for those 60 or younger and 35 months for those older than 60. Pretransplant diagnoses varied slightly between younger versus older patients but overall were comparable (Table 2).

Table 1. AGE AND PRETRANSPLANT LOCATION

Table 2. PRETRANSPLANT DIAGNOSES BY AGE

* Bonferroni multiple comparisons.

Recipient Age and Patient/Graft Survival

When reviewed alone, recipient age was linked significantly to both patient and graft survival. Decreased patient survival was shown with advancing age when patients were grouped as younger than 60 versus 60 or older, or younger than 65 versus 65 or older (Fig. 1), but not when grouped as younger than 50 versus 50 or older (data not shown). Likewise, patients younger than 60 had significantly better graft survival than those 60 or older (3-year graft survival 74.7% vs. 57.1%, respectively, P ⩽ .0001).

Figure 1. Patient survival by age group.

Recipient Age and Pretransplant Status

Kaplan-Meier survival probability curves were compared by pretransplant status or location (home, hospital, and intensive care unit). Survival for patients who were not inpatients immediately before the transplant (“home”) was not influenced by age group (Fig. 2). In fact, for patients younger than 60, pretransplant location (inpatient vs. outpatient) had a statistically significant but numerically trivial impact on survival. By contrast, patients 60 or older who were inpatients at the time of transplant had lower survival probabilities. Even worse outcomes were noted in patients 60 or older who were in the intensive care unit before the transplant.

Figure 2. (A) Patient survival by age group vs. pretransplant location (home, hospital, intensive care unit). (B) Patient survival for those younger than 60: inpatient (hospital or intensive care unit) vs. outpatient (home). (C) Patient survival by age group: inpatients (hospital or intensive care unit before the transplant).

When these comparisons were repeated using age 65 as a cutoff, differences among the age groups were accentuated. When outpatient status was compared with grouped inpatient status (hospital plus intensive care unit), the differences remained striking. Even the outpatients older than 65 had survival statistics comparable to those of the younger outpatients. When only inpatients were compared by age group, the decreased survival in advancing age groups remained highly significant.

Recipient Age and Pretransplant Laboratory Values

We used bilirubin as one pretransplant surrogate marker of the severity of liver disease, arbitrarily grouping our age-versus-bilirubin analyses with total bilirubin values of 0 to 5 mg/dL, 5 to 10 mg/dL, 10 mg/dL or more, and 20 mg/dL or more. Our intent was to identify values above which a higher risk (if any) could be predicted.

Pretransplant total bilirubin had no impact on posttransplant survival probabilities in patients younger than 60. However, patients 60 or older with a bilirubin level of 10 mg/dL or greater had a notable decrease in survival noted early after the transplant and persisting through 5 years (Fig. 3).

Figure 3. Patient survival by age group and pretransplant bilirubin level.

We used a serum albumin level of 3 g/dL to group patients and compare age-based patient outcomes. For patients younger than 60, an albumin level of less than 3 g/dL was not predictive of poorer outcome. However, at age 60 or older, an albumin level of less than 3 g/dL (compared with 3 g/dL or more) was significantly associated with poorer patient survival (Fig. 4).

Figure 4. Patient survival by age group and pretransplant albumin level.

Patients were grouped by pretransplant prothrombin time (⩽20 seconds or less vs. >20 seconds) and survival by age group was compared. For patients younger than 60, no influence on survival could be predicted by prothrombin time. However, a markedly prolonged (>20 seconds) prothrombin time in patients 60 or older was linked to a significantly poorer outcome (Fig. 5).

Figure 5. Patient survival by age group and pretransplant prothrombin time.

Child-Pugh Classification

Child-Pugh scores were tabulated for all patients and assigned values of low (score 0–5), medium (6–10), and high (>10). For patients younger than 60 at the time of transplant, no relation was found between pretransplant Child-Pugh score groups and outcome. Conversely, patients 60 or older who had a high score (>10) had noticeably poorer survival statistics (Fig. 6).

Figure 6. Patient survival by age group and Child-Pugh score.

Child-Pugh scores with the addition of a nutritional status component were also compared between age groups. Patients 60 or older with higher scores (>12) had significantly (P < .0001) lower survival probabilities (data not shown). As with the scores exclusive of nutritional assessment, the younger patients showed no decrease in survival despite a higher score (data not shown). Scores with nutritional components were assigned values of low (0–7), medium (7–12), and high (>12).

To assess the impact of pretransplant nutritional status alone (independent of Child-Pugh score) on age, groups were compared according to survival outcomes by age groups in three groups of patients: well nourished, moderately malnourished, and severely malnourished. Nutritional status was assessed before the transplant by a registered dietitian using subjective global assessment. ¹³ Patients 60 or older patients who were considered severely malnourished had poorer survival than the better-nourished patients of the same age and than the patients younger than 60, well nourished or not.

Recipient and Donor Age

A reduction in survival statistics was linked to donor age older than 50, regardless of whether recipients were in the younger or older group. Patients younger than 60 receiving a donor liver from a donor younger than 50 had approximately 10% higher survival probabilities than any other donor–recipient age combination (P < .0001). Nonetheless, no differences were noted in survival statistics between any other donor–recipient combination. This analysis was performed blind to donor–recipient gender combinations. The penalty incurred with the use of older donors has been previously reported by our group. ¹⁰ Recipients older than 60 did not receive an older donor liver more frequently than younger patients.

Cox Regression and Logistic Regression Analyses

A Cox regression analysis of all studied risk factors associated with poorer survival was performed. Age group 60 or older, inpatient or intensive care unit status before the transplant, donor age (older than 50), and bilirubin level more than 10 mg/dL were all identified as independent risk factors, with risk ratios of 1.3 to 2.3 (Table 3).

Table 3. STEPWISE COX REGRESSION ANALYSIS OF PATIENT SURVIVAL

Because our Kaplan-Meier survival curves revealed that the biggest drop in survival probability occurred in the first year after the transplant, a logistic regression analysis was performed for the 3-month and 1-year posttransplant time points. The results were similar to the Cox regression in identifying age 60 or older, hospital or intensive care unit status, donor age older than 50, and pretransplant bilirubin level more than 10 mg/dL as independent risk factors (Table 4).

Table 4. STEPWISE LOGISTIC REGRESSION TO PREDICT PATIENT DEATH

Patient Selection Bias

We examined our patient population by age group to uncover possible bias in patient selection (e.g., choosing healthier elderly patients for transplant). We questioned, for example, whether elderly (age 60 or older) recipients had similar representation as younger patients of their hepatic synthetic function. The results were mixed. Older patients tended to be less well nourished than younger ones and tended to have ascites more commonly before the transplant. However, younger patients were more likely to have a serum albumin level of less than 3 g/dL, a bilirubin level of greater than 5 mg/dL, and a prothrombin time greater than 20 seconds (Table 5).

Table 5. RECIPIENT PRETRANSPLANT CHARACTERISTICS BY AGE

DISCUSSION

This single-center study found lower survival probabilities in some recipients of liver transplants who were older than 60. Further subpopulations of this age group (inpatients, those with higher bilirubin levels, higher prothrombin times, lower albumin levels, poor nutritional status) were especially at risk for lower survival. The number of patients studied here represents by far the largest such study in the literature. We undertook this analysis because our experience seemed contrary to what has been reported in the literature: namely, other investigators had found no decrease in survival among older liver transplant recipients. ^8,12 Our definition of “elderly” as 60 years or older was derived from a statistical analysis showing no age-linked impact below that age.

The appropriateness of allocating transplanted organs to elderly patients is increasingly addressed in the literature. In kidney transplantation, for example, some authors argue firmly against age restrictions for kidney recipients. ¹⁴ A recent multicenter report of 735 liver transplant recipients (17% older than 60), although acknowledging significantly lower survivals in older patients, also enthusiastically endorsed transplantation for the elderly. ¹⁵ Regrettably, those authors made no attempt to stratify pretransplant predictors of poor outcome among their elderly population.

One question raised by this study was whether the younger and elderly patient populations were similar in illness severity. Our data suggest that the younger patients in fact had poorer hepatic synthetic function than the older ones, whereas older ones were more catabolic (see Table 5). Another question is whether causes of death were similar for younger and older patients. Although we did not focus on that question, previous reports from our center have noted that death from cardiovascular disease is more prevalent among older patients. ^16—18 Other investigators have noted that “excess mortality” in the elderly is often due to nonhepatic causes. ¹⁵

Our analysis suggests that the survival differences between younger and older patients are not due to differences in pretransplant diagnoses between the groups (see Table 2). The higher proportion of patients with fulminant or subfulminant liver failure (a “favorable” diagnosis) in the younger patients was partially offset by a higher proportion of cryptogenic cirrhosis (another “favorable” diagnosis) in the older patients. Moreover, the younger fulminant or subfulminant group represented less than 5% of the total study population.

Similarly, it is unlikely that the noted survival differences could be explained by the age ranges of our younger patients. Although very young patients (i.e., infants) are known to have slightly lower survival rates after liver transplants, our study patient population was overwhelmingly composed of patients older than 17 (see Table 1).

The pretransplant parameters studied here have inconsistently been linked to poorer patient outcome. The survival penalty of being an inpatient or in the intensive care unit before the transplant has been noted for some time. ^10,19 The physiologic status of the liver transplant recipient is an important factor in determining outcome. ²⁰ In nontransplant cases, an elevated preoperative serum bilirubin level has been associated with an adverse outcome after nonshunt laparotomies in patients with cirrhosis. ¹² The link between pretransplant bilirubin levels and outcome in transplant recipients has not previously been explored. Likewise, we are the first to note that pretransplant prothrombin time and albumin level, as markers of hepatic synthetic function, are linked to variations in outcome.

Several important conclusions can probably be drawn from our results. First, in the population of elderly liver transplant recipients, those who seem to be at risk for poorer results are those who are inpatients or in the intensive care unit before the transplant. Efforts at improving the results of liver transplantation for the elderly should be implemented. Moreover, the news is good for the elderly waiting at home for their transplant, suggesting that an age limit for that population still awaits definition. ⁶

We are in an era of bitter discord over appropriate liver allocation and distribution, where median waiting times in several parts of the country exceed 2 years and where lawsuits are filed in federal court over liver allocation policies. ²¹ Regrettably, as the waiting list grows so does the number of patients, young and old, dying for lack of a suitable organ. Until survival can be improved in some subgroups of elderly liver transplant recipients, perhaps their suitability for transplant should be seriously questioned. Our results suggest it would be prudent to forgo transplantation in the patient older than 60 who is an inpatient or in the intensive care unit, whose serum albumin level is less than 3 g/dL, whose total bilirubin level exceeds 10 mg/dL units, and whose prothrombin time is greater than 20 seconds.