Outcome of Patients with Cystic Fibrosis Awaiting Lung ... - ATS Journals

13 downloads 0 Views 163KB Size Report
Thirty-seven patients died (Died-waiting Group; n 37); 76 patients under- went bilateral ..... Hayden, A. M., R. C. Robert, J. M. Kriett, C. M. Smith, K. Nicholson, and S. W. .... ner, W. D. Ogden, J. R. Yankaskas, J. H. Westerman, J. T. Thompson,.
Outcome of Patients with Cystic Fibrosis Awaiting Lung Transplantation CARMINE DARIO VIZZA, ROGER D. YUSEN, JOHN P. LYNCH, FRANCESCO FEDELE, G. ALEXANDER PATTERSON, and ELBERT P. TRULOCK Department of Cardiology; “La Sapienza” University School of Medicine, Rome, Italy; Washington University School of Medicine, St. Louis, Missouri

Cystic fibrosis is a common indication for lung transplantation. Under the current organ allocation system, donor lungs are distributed to patients based solely on their accrued waiting time, and the death rate on the waiting list has been high. Physiologic parameters have been used to guide the referral, but risk factors for death while awaiting transplantation have not been well defined. This study aimed to identify factors at the time of evaluation that were associated with death on the waiting list. A consecutive cohort of 146 patients with cystic fibrosis who were listed for lung transplantation was retrospectively reviewed. Characteristics of patients who died awaiting transplantation were compared with those of patients who survived until transplantation or the end of the study. Thirty-seven patients died while waiting, 76 underwent transplantation, and 33 were alive and still waiting. Actuarial survival rates for the entire cohort were 81% at 1 yr, 67% at 2 yr, and 59% at 3 yr. Although a multivariate Cox proportional hazards model (␹2 ⫽ 29.6; p ⬍ 0.001) identified shorter six-minute walk distance (50 m increments; RR, 0.69; 95% CI, 0.57 to 0.84), higher systolic pulmonary artery pressure (5 mm Hg increments; RR, 1.41; 95% CI, 1.11 to 1.80), and diabetes mellitus (RR, 1.57; 95% CI, 1.06 to 2.32) as significant risk factors for death on the waiting list, these factors and other features overlapped considerably between the group of patients who died waiting and the group who lived until transplantation or the end of the study. The transplant evaluation selects a rather homogeneous cohort of patients for the waiting list. Unless outcome on the waiting list can be reliably predicted, establishing criteria to allocate donor lungs according to medical urgency may not be feasible.

In the U.S. Scientific Registry, cystic fibrosis has been an increasingly frequent indication for lung transplantation during the last decade (1). In 1997, cystic fibrosis was the underlying disease in approximately 19% of lung transplant recipients in the United States, and it was the most common reason for lung transplantation after chronic obstructive pulmonary disease (1). Despite this growing experience with lung transplantation for cystic fibrosis, identifying appropriate candidates for transplantation and timing the procedure are still complex issues. Patients who are listed for lung transplantation face a long wait. For patients who initially registered on the lung transplant waiting list in 1996, the median waiting time to transplantation was 567 d (1). Under the present organ allocation policy in the United States, donor lungs are distributed to patients on the waiting list who appropriately match the avail-

able donor for blood type and body size, solely according to their seniority, i.e., their accrued active waiting time, on the United Network for Organ Sharing (UNOS) national waiting list. Recently, the Department of Health and Human Services has advocated a revision of the organ allocation system that would also consider medical urgency (2, 3). However, factors that affect prognosis while awaiting lung transplantation have not been extensively studied, and if reliable predictors of death on the waiting list cannot be identified, prioritizing patients for transplantation by medical urgency will be impossible. In most studies of patients awaiting lung transplantation, the underlying lung disease itself has been the main determinant of mortality on the waiting list, and patients with pulmonary fibrosis, cystic fibrosis, and pulmonary hypertension have had higher death rates on the waiting list than those with chronic obstructive pulmonary disease (4–8). Global measures of functional status such as six-minute walk distance (5) and quality of well-being (9) have had some predictive value. However, in a risk assessment of all patients on a waiting list, physiologic parameters have been less discriminating because of the disparate patterns among the various diseases. Physiologic parameters, especially the FEV1, PaCO2, and PaO2, have significant ramifications for mortality in patients with cystic fibrosis (10), and guidelines for timing transplantation have incorporated these predictions (11–13). However, trends in the FEV1 and other, less quantifiable, clinical factors also have prognostic implications. Relying too heavily on the FEV1 criterion to time transplantation could adversely affect the overall survival of transplant candidates with cystic fibrosis (14, 15). The risk of death while awaiting lung transplantation has been high for patients with cystic fibrosis (4, 6, 7). At some transplant centers, 15 to 40% of patients with cystic fibrosis have died before transplantation (5, 11, 16–18). In previous studies of patients with cystic fibrosis awaiting heart-lung or lung transplantation, FEV1, PaCO2, and six-minute walk distance have been the most dependable predictors of death before transplantation (16, 17). The purpose of this study was to identify risk factors for death in patients with cystic fibrosis while awaiting transplantation and to examine the relationship between their clinical and physiologic characteristics and their outcomes on the waiting list.

METHODS Patients

(Received in original form October 25, 1999 and in revised form February 22, 2000) Supported in part by Grant K23 HL-04236-01 from the NIH Lung and Blood Institute, by National Research Service Award F32 HS00124-02 from the Agency for Health Care Policy and Research, and by the Rosemary and I. Jerome Flance Professorship in Pulmonary Medicine. Correspondence and requests for reprints should be addressed to Elbert P. Trulock, M.D., Division of Pulmonary & Critical Care Medicine (Box 8052), Washington University School of Medicine, 660 South Euclid Avenue, St. Louis, MO 63110. E-mail: [email protected] Am J Respir Crit Care Med Vol 162. pp 819–825, 2000 Internet address: www.atsjournals.org

The records of 146 consecutive patients with cystic fibrosis who were listed for lung transplantation at Barnes-Jewish Hospital between January 1, 1989 and May 12, 1998 were retrospectively reviewed. These patients had been screened and evaluated for transplantation while clinically stable. Standard criteria were used for evaluating and selecting potential recipients (11–13, 19). Patients who were infected with Burkholderia cepacia were not accepted for transplantation. Data that were collected during the evaluation for transplantation included age, sex, height, weight, medical history, and physical examination, pulmonary function, arterial blood gases, six-minute walk distance, radionuclide ventriculogram, and cardiac catheterization.

820

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

Patients were classified as having diabetes mellitus if they were being treated with insulin or an oral hypoglycemic drug at the time of their evaluation. Likewise, patients were labeled as receiving corticosteroid therapy if they were receiving prednisone or another systemically administered corticosteroid medication when they were evaluated. Inhaled corticosteroid drugs were not included in the corticosteroid therapy category.

Pulmonary Function Tests Pulmonary function tests were done with a MedGraphics System 1085 (MedGraphics Corp., St. Paul, MN) according to American Thoracic Society standards, and customary normal reference values were used for spirometry, lung volumes, and diffusing capacity (20–22). Functional residual capacity was measured by plethysmography; total lung capacity and residual volume were calculated by standard formulas. Diffusing capacity for carbon monoxide was measured by the singlebreath technique. A specimen of arterial blood was taken with the patient seated and breathing room air; pH, PaCO2, and PaO2 were measured with a blood gas analyzer (Model BG3; Instrumentation Laboratory, Lexington, MA).

Six-minute Walk Test Exercise capacity was assessed by a standardized six-minute walk test (23). Patients were asked to walk as quickly as possible, but comfortably, and were encouraged along the way. During the walk, oxygen saturation was monitored continuously with a portable pulse oximeter, and supplemental oxygen was administered and adjusted to maintain SpO2 above 89%. The distance walked in 6 min was recorded, and it was referenced to a normal value for a healthy adult (24).

Radionuclide Ventriculography Right (RVEF) and left (LVEF) ventricular ejection fractions were measured by gated equilibrium radionuclide ventriculography after a peripheral intravenous injection of 750 MBq technetium-99m human serum albumin. An ejection fraction of 45% was considered the lower limit of normal for both RVEF and LVEF.

Cardiac Catheterization Right heart catheterization was performed by standard technique. Pressures were measured from the midchest position with a fluidfilled catheter and pressure transducer; the average values over three respiratory cycles were recorded. Cardiac output was measured by the thermodilution technique (American Edwards Laboratories, Santa Ana, CA). The value for pulmonary vascular resistance index (PVRI) was calculated with the formula, PVRI ⫽ Ppa ⫺ Ppw /CI, where Ppa is the mean pulmonary artery pressure, Ppw is the mean pulmonary wedge pressure, and CI is the cardiac index. Coronary angiography was not done because of the low risk for coronary artery disease in the cohort.

Patient Management and Follow-up Soon after being evaluated and accepted, patients were registered on the UNOS national waiting list for lung transplantation. Routine medical management was continued during the waiting period under the supervision of the primary/referring physician(s) and/or the transplant physicians. Organs were allocated to patients through the UNOS Network in accordance with its policies and procedures. Throughout the period of this study, lungs were distributed to patients on the waiting list who appropriately matched the blood type and body size of the available donor, solely based on their accrued active waiting time. The status of all patients, including their date of death or transplantation, was determined at the end of the study period (February 12, 1999). Deaths on the waiting list and transplants were corroborated with UNOS records.

Data Analysis The data from the patients’ evaluations, which coincided within a few days with their registration on the UNOS waiting list, were compiled and analyzed. Only pulmonary function tests and six-minute walk tests that were done within 30 d of the initial evaluation were utilized, and only radionuclide ejection fractions and hemodynamic measure-

VOL 162

2000

TABLE 1 PROFILE OF PATIENTS WHEN EVALUATED FOR TRANSPLANTATION Tx/Alive-Waiting

General Characteristics Age, yr Sex, female Height, cm Weight, kg Ideal body weight, % Patients with Diabetes mellitus Corticosteroid therapy Selected laboratory values Blood glucose, mg/dl Serum albumin, g/dl Serum creatinine, mg/dl Hemoglobin, g/dl WBC, 103/mm3

Died-Waiting

n

Value*

n

Value*

p Value

109 109 109 109 109

28 ⫾ 7 45 (41%) 168 ⫾ 9 51.2 ⫾ 9.8 83 ⫾ 11

37 37 36 36 36

30 ⫾ 9 20 (54%) 165 ⫾ 9 48.8 ⫾ 10.5 84 ⫾ 15

0.105 0.177 0.073 0.200 0.685

106 107

31 (29%) 10 (9%)

37 37

19 (51%) 10 (27%)

0.015† 0.007†

107 106 107 101 102

131 ⫾ 75 3.9 ⫾ 0.4 0.8 ⫾ 0.2 13.3 ⫾ 1.8 11.2 ⫾ 3.8

36 36 36 36 36

172 ⫾ 95 3.8 ⫾ 0.3 0.8 ⫾ 0.2 12.2 ⫾ 1.6 13.5 ⫾ 4.5

0.020† 0.156 0.635 0.002† 0.004†

* Mean ⫾ SD or n (percentage of group). † Statistically significant difference between groups (p ⬍ 0.05).

ments that were obtained within 90 d of the initial evaluation were used. For comparative analyses, the patients were divided into two groups, those who died while awaiting transplantation (Died-waiting group) and those who either underwent transplantation or were alive and still awaiting transplantation on February 12, 1999 (Tx/Alive-waiting group). Data were obtained from chart review and were stored in a spreadsheet (Microsoft Excel; Microsoft Corp., Redmond, WA). Exploratory analyses and error-check calculated fields were used to screen the data for entry errors. After discrepancies were resolved, data were transferred to a SPSS worksheet (SPSS version 6.1 for the Macintosh; SPSS Inc., Chicago, IL) for analysis. Descriptive statistics were used to describe the patients’ characteristics. Continuous data are expressed as mean ⫾ standard deviation, and categorical data are expressed as counts and proportions. Distributions of data were determined and analyzed. Unrelated two-group comparisons were done with unpaired, two-tailed t tests for means if the data were normally distributed or with Wilcoxon’s rank-sum tests if the data were not normally distributed. Chi square or Fisher’s exact tests were used to analyze the categorical data. Actuarial survival was determined by the Life Table method. Kaplan-Meier (product-limit) graphs were used to demonstrate survival over time. Patients who underwent lung transplantation were censored at the time of their operation, and patients who were alive and still waiting for transplantation were censored on the date that the study concluded. Cox proportional hazards regression methods were used to identify risk factors for death while on the transplant wait list and to determine the association among baseline patient characteristics and outcomes. Time to death while awaiting transplantation was selected as the primary outcome. Univariate proportional hazards analyses were performed, and Wald chi-square p values were calculated. The likelihood ratio method was used to determine hazard ratios, and the hazard ratio was used to approximate the relative risk (RR). After all Cox univariate analyses were performed, the covariates were inversely ordered by p value (smallest to largest). Because of the large number of variables that were being assessed, a strict univariate p-value criterion (p ⬍ 0.01) was used to select those variables that were initially entered into the model. This approach was taken to minimize the probability that the value of the calculated test statistic, or an extreme value, occurred by chance. Covariates with p ⬍ 0.01 were entered into a forward stepwise multivariate Cox regression analysis. A p value ⬎ 0.10 was the criterion to remove covariates from the model. Model and covariate characteristics were assessed at each step, and the most parsimonious model was sought. The analysis was done manually, but an automated analysis yielded the same result. After a final model was achieved, other covariates that were considered potentially clinically relevant, or that exhibited an impressive

Vizza, Yusen, Lynch, et al.: Patients with Cystic Fibrosis Awaiting Lung Transplantation

821

TABLE 2 CARDIOPULMONARY FEATURES OF PATIENTS WHEN EVALUATED FOR TRANSPLANTATION Tx/Alive-Waiting n Ejection fractions Right ventricle, % Left ventricle, % Hemodynamics Pressures, mm Hg RA PA systolic PA diastolic PA mean Wedge Cardiac index, L/min/m2 PVRI, wood unit, m2 Pulmonary function Spirometry‡ FVC, L FVC, % FEV1, L FEV1, % FEV1,/FVC ⌬FEV1, % Patients with ⌬FEV1, ⬎ 15% Lung volumes§ TLC, L TLC, % RV, L RV, % RV/TLC Diffusing capacity DLCO, ml/min/mm Hg DLCO, % Gas exchange¶ PaCO2, mm Hg PaO2, mm Hg AaPO2 Patients with PaO2 ⬍ 60 mm Hg Exercise tolerance Six-minute walk, m Six-minute walk, %

Died-Waiting

Value*

n

Value*

105 106

43 ⫾ 8 61 ⫾ 7

36 36

40 ⫾ 11 61 ⫾ 9

0.250 0.828

99 101 101 97 100 99 95

4.7 ⫾ 2.9 36 ⫾ 7 19 ⫾ 6 26 ⫾ 5 9⫾4 3.9 ⫾ 0.7 4.2 ⫾ 1.7

30 32 32 32 32 31 31

3.7 ⫾ 3.3 39 ⫾ 10 20 ⫾ 7 28 ⫾ 7 9⫾4 4.1 ⫾ 0.8 4.6 ⫾ 1.9

0.102 0.143 0.520 0.154 0.582 0.205 0.256

108 1.73 ⫾ 0.60 36 1.46 ⫾ 0.64 108 40 ⫾ 11 36 36 ⫾ 12 108 0.78 ⫾ 0.24 36 0.74 ⫾ 0.26 108 22 ⫾ 6 30 22 ⫾ 5 108 0.49 ⫾ 0.12 36 0.53 ⫾ 0.12 90 10 ⫾ 12 30 8⫾7 90 19 (21%) 30 5 (17%) 81 6.4 ⫾ 1.7 81 109 ⫾ 21 81 4.4 ⫾ 1.4 81 313 ⫾ 85 81 0.68 ⫾ 0.08 101 15.3 ⫾ 5.6 101 43 ⫾ 14 103 103 103 103

47 ⫾ 11 57 ⫾ 12 34 ⫾ 9 52 (50%)

101 396 ⫾ 96 101 53 ⫾ 13

p Value



0.006 0.031† 0.251 0.823 0.011† 0.687 0.598

31 5.7 ⫾ 1.5 31 101 ⫾ 19 31 4.1 ⫾ 1.3 31 287 ⫾ 74 31 0.70 ⫾ 0.09

0.053 0.082 0.238 0.132 0.251

32 11.9 ⫾ 5.1 32 35 ⫾ 12

0.001† 0.005†

46 ⫾ 8 52 ⫾ 11 41 ⫾ 10 25 (71%)

0.513 0.008† ⬍ 0.001† 0.031†

31 343 ⫾ 97 31 47 ⫾ 14

0.012† 0.046†

35 35 35 35

Definition of abbreviations: ⌬FEV1 ⫽ change in FEV1, postbronchodilator versus prebronchodilator; PA ⫽ pulmonary artery; PVRI ⫽ pulmonary vascular resistance index; RA ⫽ right atrium. * Mean ⫾ SD or n (percentage of group). † Statistically significant difference (p ⬍ 0.05). ‡ Prebronchodilator values, except ⌬FEV1. § Plethysmographic measurements. ¶ Room air; at rest.

hazard ratio, were forced into the model to appraise their effect, but model goodness-of-fit did not significantly improve. Time-dependent covariates (interaction terms) and “minus log–minus log” plots were used to confirm that the proportional hazards assumption was met. Regression diagnostics, including tests for interactions, outliers and influential points, were performed, but the final model remained unchanged. Subgroup analyses were not performed because of the size of the sample and the number of outcome events. To assess the potential impact of missing data, worst-case scenario substitutions for the missing information were performed for both of the final model’s continuous variables (six-minute walk distance and systolic pulmonary artery pressure). This was not done for the categorical variable (diabetes mellitus) because the diabetic status was unknown for only three patients. The missing data for six-minute walk distance and pulmonary artery systolic pressure in the Tx/alivewaiting Group were replaced by the median values of these same variables from the Died-waiting Group, and the missing data in the Died-waiting Group were replaced by median values from the Tx/ alive-waiting Group. The model that resulted from ascribing these values to the missing data was not significantly different (␹2 ⫽ 25; p ⬍

Figure 1. Kaplan-Meier survival plot of the proportion of patients with cystic fibrosis remaining alive while awaiting lung transplantation. Recipients were censored at the time transplantation. The number of patients available for analysis at each 6-mo interval is shown below the graph. Vertical hatch marks indicate censored cases.

0.001) from the original model. Even with the worst-case substitutions for the missing data, the six-minute walk distance and pulmonary artery systolic pressure still had statistically significant p values, but the relative risks were less extreme.

RESULTS Follow-up was complete, and outcome was determined for all 146 patients through the end of the study period. Thirty-seven patients died (Died-waiting Group; n ⫽ 37); 76 patients underwent bilateral lung transplantation, and 33 patients were alive and still waiting at the end of the study period (Tx/alive-waiting Group; n ⫽ 109). Profiles of the patients’ characteristics when they were listed for transplantation are presented in Tables 1 and 2. As expected, in advanced cystic fibrosis, the patients were relatively young, and their body weight was below the norm. A modest number of the patients had diabetes mellitus, and some were receiving systemic corticosteroid therapy. Consistent with their underlying advanced bronchiectasis, the patients had severe obstructive lung disease, impaired oxygenation, and poor exercise tolerance. However, their hemodynamic profiles were rather unremarkable. Although there were some outliers, the mean pulmonary artery pressure was at the upper limit of the normal range, the mean cardiac index was normal, and the mean right ventricular ejection fraction was slightly below the lower limit of normal. Some significant differences were apparent between the patients who died while awaiting transplantation and those who survived to transplantation or were alive and still waiting at the end of the study (Tables 1 and 2). In the Died-waiting Group, corticosteroid treatment and diabetes mellitus were more prevalent, and the fasting blood glucose was higher, than in the Tx/alive-waiting Group. In addition, the hemoglobin was lower, and the white blood cell count higher, in the Diedwaiting Group. The cardiopulmonary parameters differed significantly between the two groups in only a few categories (Table 2). The hemodynamic profiles and the ventricular ejection fractions were similar. Among the pulmonary function measurements, only the vital capacity and the diffusing capacity were significantly lower in the Died-waiting Group. Comparisons of prebroncholator and postbronchodilator spirometric values yielded

822

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE TABLE 3

UNIVARIATE COX PROPORTIONAL HAZARDS ANALYSIS OF RISK FACTORS FOR DEATH WHILE AWAITING TRANSPLANTATION Variable* (Increment)† Six-minute walk, 50 m WBC, 103/mm3 DLCO, 1 ml/min/mm Hg ‡ AaPO2, 5 mm Hg PaO2, 5 mm Hg‡ DLCO, % (5%) Diabetes mellitus Hemoglobin, 1 g/dl Six-minute walk, % (5%) Blood glucose, 10 mg/dl Systemic corticosteroid use PaO2 ⬍ 60 mm Hg‡ Prebronchodilator FVC, 0.250 L PA systolic pressure, 5 mm Hg

n

Relative Risk

95% CI

p Value

132 138 133 138 138 133 143 137 132 143 144 138 144 133

0.73 1.13 0.89 1.31 0.81 0.81 1.69 0.74 0.82 1.04 1.68 1.71 0.78 1.35

0.62–0.87 1.05–1.21 0.83–0.95 1.11–1.54 0.71–0.92 0.71–0.92 1.22–2.34 0.60–0.89 0.72–0.94 1.01–1.07 1.16–2.41 1.18–2.48 0.66–0.93 1.08–1.68

⬍ 0.001 ⬍ 0.001 0.001 0.001 0.001 0.001 0.002 0.002 0.003 0.004 0.005 0.005 0.005 0.008

* Variables with p ⬍ 0.01. † Increment in the continuous variable used to interpret the relative risk. ‡ Room air; at rest.

similar results. Minimal, comparable hypercapnia was present in both groups, but hypoxemia was more severe and more prevalent in the patients who died while waiting. Lastly, exercise tolerance, as gauged by the six-minute walk distance, was more limited in the Died-waiting Group. Kaplan-Meier survival from the time of evaluation for transplantation is shown in Figure 1. Actuarial survival rates for the entire cohort of 146 patients, with recipients censored at the time of transplantation, were 81% at 1 yr, 67% at 2 yr, and 59% at 3 yr. Mean survival in the Died-waiting Group was 7.4 ⫾ 5.8 mo. In the Tx/alive-waiting Group, the average time to transplantation for the recipients was 14.3 ⫾ 9.9 mo, and the waiting time for the patients who were alive and still waiting was 21.6 ⫾ 11.8 mo. Because of the long inherent waiting time for transplantation, patients who did not reside in the vicinity of our medical center returned home and remained under the care of their primary/referring physician(s) while accruing some seniority on the waiting list. Therefore, many of the deaths on the waiting list did not occur at our institution, and the cause of death was not available in the majority of cases. The mode of death was known in 10 patients, and respiratory complications were the main contributor to all of these fatalities. The results of the univariate proportional hazards assessment of risk factors for death while awaiting transplantation are exhibited in Table 3. Among the patients’ characteristics, diabetes mellitus, corticosteroid therapy, and higher blood glucose were significantly associated with an increased risk of dying while waiting for transplantation. Several parameters of lung function and oxygenation were also related to the risk of death. These included a lower FVC; a lower PaO2, especially less than 60 mm Hg; a larger AaPO2; and a lower DLCO. The

DISCUSSION The success of lung transplantation for a variety of end-stage lung diseases has fostered an increasing demand. Registrations on the UNOS national waiting list for lung transplantation increased from 308 at the end of 1990 to 3,165 by the end of 1998, and from 1995 to 1998 the size of the list grew 15 to 20% each year (1). The demand continues to exceed the supply of suitable donor organs. Consequently, the median waiting time for lung transplantation has almost tripled from 213 d for registrants in 1990 to 567 days for those in 1996 (25), and, recently, approximately 15% of patients on the lung waiting list have died each year while awaiting transplantation (1). Cystic fibrosis is now the second leading indication for lung or heart-lung transplantation (26), and the results have been good (25, 27, 28). However, the risk of death while awaiting lung transplantation has been high for patients with cystic fibrosis. At some transplant centers, 15 to 40% of patients with cystic fibrosis have died before transplantation (5, 11, 16–18). Under the UNOS organ allocation policy for lung transplantation, donor lungs have been and still are distributed to matching patients solely on the basis of their accrued waiting time. Although the allocation algorithm may be revised in upcoming years in response to the recent mandate from the Department of Health and Human Services, there has been no priority for medical urgency.

TABLE 4

Six-minute walk (50 m increments) Pulmonary artery systolic pressure (5 mm Hg increments) Diabetes mellitus

2000

only hemodynamic component that appeared relevant in the univariate analysis was the pulmonary artery systolic pressure, and a higher pressure increased the risk of death. The six-minute walk distance was strongly and inversely correlated with the risk of death before transplantation. In univariate analyses of the patients’ time of listing and sequential case number on the waiting list, significant temporal trends were not identified. The results of the multivariate analysis are presented in Table 4. Three variables, the six-minute walk distance, the systolic pulmonary artery pressure, and the presence of diabetes mellitus, emerged as significant independent risk factors for death on the waiting list in the final model after adjusting for confounders (p ⬍ 0.001). For each 50 m increase in the sixminute walk distance, the risk of death decreased by 31%, and, for each 5 mm Hg increase in the systolic pulmonary artery pressure, the chance of death increased by 41%. Finally, the presence of diabetes mellitus increased the risk of death before transplantation by 57% (Figure 2). Box plots of the six-minute walk and the pulmonary artery systolic pressure data are displayed in Figures 3 and 4, respectively. Although the mean six-minute walk distance was significantly different between the Died-waiting and the Tx/Alivewaiting groups and was a significant independent predictor of death in the multivariate analysis, the walk distances of the two groups overlapped considerably, and no cutoff distance for the 6-minute walk distinctly separated the two groups. Similarly, there was no systolic pulmonary artery pressure that clearly divided the two groups.

MULTIVARIATE COX PROPORTIONAL HAZARDS MODEL OF RISK FACTORS FOR DEATH WHILE AWAITING TRANSPLANTATION* Variable

VOL 162

Relative Risk

95% CI

p Value

0.69 1.41 1.57

0.57–0.84 1.11–1.80 1.06–2.32

⬍ 0.001 0.005 0.023

* Model chi-square ⫽ 29.6; p ⬍ 0.001; n ⫽ 119 (patients with data available for all three variables).

Vizza, Yusen, Lynch, et al.: Patients with Cystic Fibrosis Awaiting Lung Transplantation

Figure 2. Adjusted Kaplan-Meier survival plot demonstrating the proportion of diabetic and nondiabetic patients with cystic fibrosis remaining alive while awaiting lung transplantation. The number of diabetic and non-diabetic patients available for analysis at each 6-mo interval is shown below the graph. The proportion of patients surviving was greater in nondiabetics than in diabetics throughout the entire period. The analysis was stratified by the presence or absence of diabetes mellitus at the time of evaluation for transplantation after adjusting for the mean values of the other two independent predictors in the final model, systolic pulmonary artery pressure and six-minute walk distance.

Determinants of outcome on the waiting list have not been delineated. Studies of institutional transplant waiting lists have suggested that the primary diagnosis and the functional status of transplant candidates are predictors of survival while waiting (4–6, 9), and an analysis of patients with emphysema, cystic fibrosis, and idiopathic pulmonary fibrosis on the UNOS waiting list in the years 1992 through 1994 confirmed that survival while awaiting transplantation was lowest in the cohort with cystic fibrosis (7). However, whether patients with a specific

Figure 3. Box plots of six-minute walk distance at transplant evaluation for patients who died while waiting (Died-waiting) and for those who survived until transplantation or were alive and still waiting at the end of the study (Tx/Alive-waiting). Boxes encompass the interquartile range of values (25th–75th percentile), and the horizontal line in the box marks the median. Whiskers demarcate the largest and smallest actual data points that are within 1.5 box lengths of the upper and lower boundaries, respectively, of the box. Diamonds depict individual outlier measurements (values between 1.5 and 3 box lengths from boundary of the box).

823

Figure 4. Box plots of pulmonary artery systolic pressure at transplant evaluation for patients who died while waiting (Died-waiting) and for those who survived until transplantation or were alive and still waiting at the end of the study (Tx/Alive-waiting). Box encompasses the interquartile range of values (25th–75th percentile), and the horizontal line in the box marks the median. Whiskers demarcate the largest and smallest actual data points that are within 1.5 box lengths of the upper and lower boundaries, respectively, of the box. Diamonds depict individual outlier measurements (values between 1.5 and 3 box lengths from boundary of the box); an asterisk represents an extreme case (value greater than 3 box lengths from the boundary of the box).

diagnosis such cystic fibrosis who were selected for transplantation because of the physiologic severity of their disease and their limited life expectancy, can be meaningfully stratified further is uncertain. This study retrospectively examined a consecutive case series of patients with cystic fibrosis awaiting lung transplantation at our center and analyzed the relationship between the patients’ clinical status when initially listed and their subsequent outcome. Among a variety of baseline clinical and physiologic variables, relatively few significant differences were found between the patients who died awaiting transplantation and those who survived to transplantation or who were alive and still waiting at the end of the study (Tables 1 and 2). Although the differences in these few parameters were statistically significant, the disparities were small and were not very impressive from the clinical perspective. The multivariate proportional hazards analysis (Table 4) revealed only three significant independent risk factors for death while waiting for transplantation—a shorter six-minute walk distance, a higher pulmonary artery systolic pressure, and the presence of diabetes mellitus. The six-minute walk test integrates cardiopulmonary function, strength, and endurance, and it has been proposed as a useful guide in the decision to list patients for transplantation (5). The six-minute walk distance was significantly shorter in patients with cystic fibrosis who died before transplantation at the Toronto Lung Transplant Program (5), but the twelve-minute walk distance was not predictive of death on the waiting list in a multivariate analysis of patients with cystic fibrosis at the Transplant Unit at Papworth Hospital (16). In our analysis the risk of death decreased noticeably as the six-minute walk distance increased, but, as shown in Figure 3, no six-minute walk distance differentiated the patients who died while waiting from those who survived to transplantation or until the end of the study period. The prevalence and prognostic implications of pulmonary hypertension have not been extensively studied in cystic fibrosis. However, a recent study discovered subclinical pulmonary

824

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

hypertension (pulmonary systolic pressure ⬎ 35 mm Hg by Doppler echocardiography) in 40% of a cohort of patients with stable cystic fibrosis and severe lung disease (mean FEV1 ⫽ 28% of predicted normal value), and survival or time to transplantation was significantly shorter in the patients with pulmonary hypertension than in those without it (29). Our analysis confirms the impression that rising pulmonary artery systolic pressure is associated with an increasing risk of death. However, there was no threshold for pulmonary artery systolic pressure that separated the Died-waiting Group from the Tx/ Alive-waiting Group (Figure 4). The impact of diabetes mellitus on the course of cystic fibrosis is unresolved. Some studies have suggested that insulindependent diabetes mellitus foreshortens life expectancy, but others have concluded that it has no effect on pulmonary function, nutritional status, or survival (30, 31). Moreover, in a recent analysis of the risk of death in patients with cystic fibrosis and advanced lung disease, diabetes mellitus was not a significant determinant (14). In our study, none of the patients had overt diabetic complications, and none was known to have died because of diabetes mellitus. Thus, the contribution of diabetes mellitus to the poor outcome cannot be readily explained. An adverse effect of diabetes mellitus on the course and control of the chronic pulmonary infection is a plausible conjecture, but a definitive answer cannot be discerned. Our study has several potential shortcomings. Errors could have been introduced during retrospective data collection, but various techniques were used to minimize errors. Some pertinent variables may not have been assessed; however, all parameters that have shown prognostic importance in previous studies and all available data from the evaluation that seemed relevant were included. Missing data could be a source of bias; however, the amount of missing information was small, and the pattern was not very different between the two groups. In addition, statistical adjustments for missing data did not significantly alter our findings. Finally, the large number of variables that were analyzed increased the possibility that a variable could appear statistically significant by chance alone, but the set of variables in the Cox multivariate analysis was limited by strict p values and by a priori hypotheses. In conclusion, three factors—a shorter six-minute walk distance, a higher systolic pulmonary artery pressure, and the presence of diabetes mellitus—were independently associated with an increased risk of death while awaiting transplantation. However, these factors and other features overlapped considerably between the two groups of patients, and no single clinical characteristic or physiologic parameter at the time of listing for transplantation reliably segregated those who died while awaiting transplantation from those who lived until transplantation or were alive and still waiting at the end of the study. Because this was a single center study, the results cannot be extrapolated to other settings without further validation. However, the recipient evaluation process is similar at most transplant centers, and it selects a rather homogeneous cohort of patients for the waiting list. Hence, this study may provide some insight into difficulties and limitations that will be encountered in trying to devise legitimate medical urgency criteria for allocating donor lungs. References 1. UNOS data. http://www.unos.org. Accessed June 8, 1999. 2. Organ Procurement and Transplantation Network; final rule. 1998. 42 CFR Part 121. Federal Register 63(No. 63):16296–16338. 3. Organ Procurement and Transplantation Network; final rule. 1999. 42 CFR Part 121. Federal Register 64(No. 202):56650–56661. 4. Hayden, A. M., R. C. Robert, J. M. Kriett, C. M. Smith, K. Nicholson,

5.

6.

7.

8.

9.

10.

11. 12.

13.

14. 15.

16.

17.

18.

19. 20.

21.

22.

23.

24.

25.

26. 27.

28.

VOL 162

2000

and S. W. Jamieson. 1993. Primary diagnosis predicts prognosis of lung transplant candidates. Transplantation 55:1048–1050. Kadikar, A., J. Maurer, and S. Kesten. 1997. The six-minute walk test: a guide to assessment for lung transplantation. J. Heart Lung Transplant. 16:313–319. D’Armini, A. M., G. Callegari, P. Vitulo, C. Klersy, M. Rinaldi, C. Pederzolli, A. M. Grande, C. Fracchia, and M. Vigano. 1998. Risk factors for early death in patients awaiting heart-lung or lung transplantation. Transplantation 66:123–127. Hosenpud, J. D., L. E. Bennett, B. M. Keck, E. B. Edwards, and R. J. Novick. 1998. Effect of diagnosis on survival benefit of lung transplantation for end-stage lung disease. Lancet 351:24–27. Geertsma, A., E. M. TenVergert, G. J. Bonsel, W. J. de Boer, and W. van der Bij. 1998. Does lung transplantation prolong life? A comparison with and without transplantation. J. Heart Lung Transplant. 17:511– 516. Squier, H. C., A. L. Ries, R. M. Kaplan, L. M. Prewitt, C. M. Smith, J. M. Kriett, and S. W. Jamieson. 1995. Quality of well-being predicts survival in lung transplantation candidates. Am. J. Respir. Crit. Care Med. 152:2032–2036. Kerem, E., J. Reisman, M. Corey, G. J. Canny, and H. Levison. 1992. Prediction of mortality in patients with cystic fibrosis. N. Engl. J. Med. 326:1187–1191. Kotloff, R. M., and J. B. Zuckerman. 1996. Lung transplantation for cystic fibrosis: special considerations. Chest 109:787–798. Yankaskas, J. R., G. B. Mallory, Jr., and C. Committee. 1998. Lung Transplantation in cystic fibrosis: consensus conference statement. Chest 113: 217–226. ASTP/ATS/ERS/ISHLT. 1998. International guidelines for the selection of lung transplant candidates. Joint statement of the American Society for Transplant Physicians (ASTP)/American Thoracic Society (ATS)/European Respiratory Society (ERS)/International Society for Heart and Lung Transplantation (ISHLT). Am. J. Respir. Crit. Care Med. 158:335–339. Milla, C. E., and W. J. Warwick. 1998. Risk of death in cystic fibrosis patients with severely compromised lung function. Chest 113:1230–1234. Doershuk, C., and R. C. Stern. 1999. Timing of referral for lung transplantation for cystic fibrosis. Overemphasis on FEV1 may adversely affect overall survival. Chest 115:782–787. Sharples, L., T. Hathaway, C. Dennis, N. Caine, T. Higenbottam, and J. Wallwork. 1993. Prognosis of patients with cystic fibrosis awaiting heart and lung transplantation. J. Heart Lung Transplant. 12:669–674. Ciriaco, P., T. M. Egan, E. L. Cairns, J. T. Thompson, F. C. Detterbeck, and L. J. Paradowski. 1995. Analysis of cystic fibrosis referrals for lung transplantation. Chest 107:1323–1327. Mannes, G. P. M., W. J. de Boer, W. van der Bij, R. G. Grevnik, G. H. Koëter, and G. L. T. Group. 1996. Three hundred patients referred for lung transplantation: experiences of the Dutch Lung Transplantation Program. Chest 109:408–413. Trulock, E. P. 1997. Lung transplantation. State of the art. Am. J. Respir. Crit. Care Med. 155:789–818. Crapo, R. O., and A. H. Morris. 1981. Standardized single breath normal values for carbon monoxide diffusing capacity. Am. Rev. Respir. Dis. 123:185–189. Crapo, R. O., A. H. Morris, P. D. Clayton, and C. R. Nixon. 1982. Lung volumes in healthy nonsmoking adults. Bull. Eur. Physiopathol. Respir. 18:419–425. Knudson, R. J., M. D. Lebowitz, C. J. Holberg, and B. Burrows. 1983. Changes in the normal expiratory flow-volume curve with growth and aging. Am. Rev. Respir. Dis. 127:725–734. Butland, R. J. A., J. Pang, E. R. Gross, A. A. Woodcock, and D. M. Geddes. 1982. Two-, six-, and 12-minute walking tests in respiratory disease. Br. Med. J. 284:1607–1608. Enright, P. L., and D. L. Sherrill. 1998. Reference equations for the sixminute walk in healthy adults. Am. J. Respir. Crit. Care Med. 158:1384– 1387. U.S. Scientific Registry for Transplant Recipients and the Organ Procurement and Transplantation Network. 1998. Annual Transplant Data Report. 1988–1997. International Society for Heart and Lung Transplantation 1999. Sixteenth Annual Data Report. http://www.ishlt.org. June 8, 1999. Egan, T. M., F. C. Detterbeck, M. R. Mill, L. J. Paradowski, R. P. Lackner, W. D. Ogden, J. R. Yankaskas, J. H. Westerman, J. T. Thompson, M. A. Weiner, E. L. Cairns, and B. R. Wilcox. 1995. Improved results of lung transplantation for patients with cystic fibrosis. J. Thorac. Cardiovasc. Surg. 109:224–235. Mendeloff, E. N., C. B. Huddleston, G. B. Mallory, E. P. Trulock, A. H.

Vizza, Yusen, Lynch, et al.: Patients with Cystic Fibrosis Awaiting Lung Transplantation Cohen, S. C. Sweet, J. Lynch, S. Sundaresan, J. D. Cooper, and G. A. Patterson. 1998. Pediatric and adult lung transplantation for cystic fibrosis. J. Thorac. Cardiovasc. Surg. 115:404–414. 29. Fraser, K. L., D. E. Tullis, Z. Sasson, R. H. Hyland, K. S. Thornley, and P. J. Hanly. 1999. Pulmonary hypertension and cardiac function in adult cystic fibrosis. Chest 115:1321–1328.

825

30. Reisman, J., M. Corey, G. Canny, and H. Levison. 1990. Diabetes mellitus in patients with cystic fibrosis: effect on survival. Pediatrics 86: 374–377. 31. Finkelstein, S. M., C. L. Wielinski, G. R. Elliott, W. J. Warwick, J. Barbosa, S.-C. Wu, and D. J. Klein. 1988. Diabetes mellitus associated with cystic fibrosis. J. Pediatr. 112:373–377.