Long-term Survival from Respiratory Failure after Marrow Transplantation for Malignancy1-3

STEPHEN

w.

CRAWFORD and FINN BO PETERSEN

Introduction

The morbidity of marrow transplantation for the treatment of malignant disorders is dependent primarily on the development of graft-versus-host disease reactions in recipients of allogeneic marrows and of conditioning-regimen-related toxicities (1, 2). Pneumonia occurs in 40 to 60010 of cases as a consequence of the intensive chemoirradiation employed in preparing a patient for marrow transplantation (3).Causes for these complications include pulmonary toxicities directly associated with the preparative regimen, infections asociated with the resultant profound immune suppression, and/or pulmonary hemorrhage associated with profound thrombocytopenia. In previous studies, the outcome of respiratory failure requiring mechanical ventilatory support has been very poor (4, 5). Approximately 5% of marrow recipients survived to 6 months after transplantation (4). In addition, intensive care for marrow recipients with respiratory failure utilizes inordinate medical resources. In a study of 50 patients by Denardo and coworkers (5), nonsurvivors of respiratory failure utilized the vast majority of blood products administered in the intensive care unit. We previously reported that patients at risk to develop these complications can be predicted. The risk factors present at time of transplantation for subsequent respiratory failure were noted to be receipt of a HLA-non-identical donor marrow, active phase of malignancy, and older age (> 21 years) (4). It is important that transplant units and patients have adequate information to assess the risks associated with marrow transplantation. Such information is crucial to discussions of advanced-care directives and cost containment. Previous studies werelimited in size and difficult to use for pretransplant counseling and informed patient choices because of the large confidence intervals that accompany the data. Accurate assessment of the validityof the previously reported risk 510

SUMMARY Respiratory failure wes measured as the institution of assisted mechanical ventilation for hypoxemic (oxygenation) or hypercarblc (ventilatory) failure after marrow transplantation. There were 348 (23%) marrow recipients who reqUired mechanical ventilation for an average of 8 days (median, 5; range, 1 to 45). The average onset of mechanical ventilation wes 39 days (median, 22; range, 0 to 172)after transplantation. Factors previously found to be associated with mechanical ventilation were tested In a Cox proportional hazards model. Older age, active malignancy at time of transplantation, and donor-recipient marrow HLA-non-Identlty were Independent rlslal for assisted mechanical ventilation after marrow transplantation. lWenty-one percent (n • 72) of the marrow recipients receiving asslated mechanical ventilation for respiratory failure were extubated. Four percent (n • 15)were discharged from the hospital, and 3% (n • 10)survived 6 months after transplantation. All of these survivors were physically functional. Three hed mild chronic respiratory symptoms and restrictive or obstructive airflow defects 1 yr after transplantation. Respiratory failure requiring assisted mechanical ventilatory support occurs In 23% of marrow recipients and Is associated with functional survival at 6 months In 3%. Older age, active malignancy at time of transplantation, and donor-recipient marrow HLA-non.ldentlty are risk factors for sUbsequent respiratory failure. In view of the poor prognosis associated with respiratory failure, these factors should be considered when counseling patients and families regarding this mode of treatment. AM REV RESPIR DIS 1992; 145:510-514

factors and further evaluation of the long-term outcomes of marrow recipients who receivemechanical ventilation would be valuable in decision-making prior to transplantation. The purpose of this study was to reassess the risk factors associated with respiratory failure after marrow transplantation for malignancy and the status of long-term survivorsof respiratory failure. Methods Subjects All patients who receiveda first marrow transplantation for malignancy between January 1986 and July 1990 wereincluded in the study (n = 1,482). Marrow Transplantation Marrow transplantation was performed according to published techniques. All research protocols wereapproved by local Institutional Review Boards. Patients received autologous, syngeneic or HLA-identical or HLAnon-identical allogeneic marrow. Patients receiving HLA-non-identical marrow shared one haplotype and differed from one to three antigens at HLA-A, B, or D loci on the unshared haplotype (6-8). Conditioning regimens for transplantation have been described in detail (8-10). Most patients receiveda combination of radiation therapy and chemother-

apy to eliminate malignant cells. Total body irradiation (TBI) was usually delivered as 12 to 15.75 Oy midline dose from opposing cobalt-60 sources given at a rate of 6 to 7 cOy/min in 6 to 12 fractions without lung shielding. Methotrexate, cyclosporine, or the combination with or without methylprednisolone were given as graft-versus-host disease (OVHD) prophylaxis (11-13). The grading of acute OVHD and the details of supportive care are described elsewhere(8-10). The transplant was done either in a conventional hospital room or in a laminar air-flow (LAF) isolation room (14). Patients not allergic to sulfamethoxazole and trimethoprim received these drugs prophylactically for 2 wk before transplant and from the time of primary discharge after transplant until normalization

(Received in original form April 22, 1991 and in revised form August 15, 1991) 1 From the Clinical Research Division, Fred Hutchinson Cancer Research Center, Seattle, Washington. • Supported by Public Health ServiceGrants CA18029and CA-47748 from the National Cancer Institute and by Grant HL-36444 from the National Heart, Lung, and Blood Institute. 3 Correspondence and requests for reprints should be adressed to Stephen Crawford, M.D., Program in Pulmonary & Critical Care Medicine, Fred Hutchinson Cancer Research Center, 1124Columbia Avenue, Seattle, WA 98104.

511

SURVIVAL FROM RESPIRATORY FAIWRE AFTER MARROW TRANSPLANTATION

of peripheral white blood cell counts or for at least 6 months (15). All data wereretrieved from clinical information prospectively collected at the time of transplantation on all patients and reviewed retrospectively for the purpose of this study. Mechanical ventilation was defined as the administration of machine-assisted tidal breaths through an endotracheal tube for hypoxemic (oxygenation)or hypercarbic (ventilatory) failure. This definition was modified slightly from our previous report to exclude patients receiving mechanical ventilation for less than 24 hours after an operative procedure or after grand mal seizure (n = 34) (5).

Pulmonary Function Testing Testing was done using American Thoracic Society guidelines (16). A Gould 1000IV pulmonary function analyzer (Gould, Inc., Dayton, OH) was used with nitrogen washout to determine lung volumes. The diffusing capacity for carbon monoxide (OLeo)was measured by the single-breath technique. Results were compared with published normal values (17-19). The actual FEV1/FVC ratio (not the percent predicted) was used as a measure of airflow since this value has been shown to be a sensitive and specific indication of airway obstruction (20). Arterial blood gas determinations were obtained by percutaneous radial artery puncture with the subject in a seated position. The AaPoz was calculated using the alveolar gas equation, PIo z - (Pao z Paco./O.8), assuming PIoz = 150 mm Hg,

Patents were censored from the analyses at death. First a stepwise, step-up Cox model was constructed using other pretransplant variables; variables were entered into the model as long as the added variable was statistically significant (p value to enter, ~ 0.05). The variables considered were indicator variables for diagnosis, whether the transplant was autologous or allogeneic, whether the patient received total body irradiation, whether the patient received a busulfan-cyclophosphamide regimen, and sex.Totalbody irradiation, radiation dose, and primary disease of Hodgkin's disease werefound to be significantlyassociated with respiratory failure in this model. Then the variables in question (age, malignancy activity status, and donor-recipient HLA-identity) were considered in a stepwise fashion for the possibilityof adding even more predictive information. These variables entered the model if they added statistically significant information.

ed in table 1 stratified according to requirement for mechanical ventilation. Twenty-three percent (n = 348) of marrow recipients received mechanical ventilatory support. The duration of support ranged from zero to 45 days (mean = 8 days) and was instituted from zero to 172days after transplantation (mean = 39 days). There appeared to be a higher proportion of patients receivingmechanical ventilation who werein a state of active malignancy (62 versus 48%) and a lower proportion who received transplants from HLA-identical marrow donors (58 versus 72%) (table I),

Statistical Analysis of Risk Factors In order to test the validity of age, malignant activity status, and donor-recipient HLA-identityas risk factors for respiratory failure, we first developed a Cox proportional hazard regression model Results that controlled for possible confounding covariates. Other factors present at time Patient Characteristics Between January 1986 and July 1990, of transplantation were entered into the 1,482patients underwent a first marrow model with mechanical ventilation as the transplant for treatment of malignancy end point. Total body irradiation, radiaat the Fred Hutchinson Cancer Research tion dose, and Hodgkin's disease as an Center. Patients ranged in age from 7 indicator variable for primary disease months to 63 yr (mean age, 29 yr), More were found to be significantly associatthan 750/0 of these patients were treated ed with respiratory failure in the model for leukemia, and 82% received al- in the absence of HLA-identity, maliglogeneic marrow transplants. Fifty-one nant activity status, and age. Next, using Statistics a model that controlled for these variStatistical analyses wereperformed testing the percent were in a state of malignant re- ables, age, malignant activity status, and of transplantation. Total lapse at time end point of respiratory failure (mechanical donor-recipient HLA-identity wereaddventilation) in a Cox proportional hazard body irradiation wasused in the pretransed in a stepwise fashion. AIl were found regression model with age, malignancy activ- plant conditioning regimen in 81 %. to add additional predictive information The patient characteristics are presentity status, and donor-recipient HLA-identity. to the model (table 2), When the reverse model was attempted, that is, age, malignant activity status, and donor-recipient TABLE 1 HLA-identity were controlled, and total PATIENT CHARACTERISTICS body irradiation, radiation dose, and primary disease of Hodgkin's disease were Respiratory Failure entered, only Hodgkin's diseaseimproved Yes Factor No the predictive power of the model (data Patients. n Age, median yr (range) Sex, Oro male Allogeneic marrow, % Diagnosis Acute lymphoblastic leukemia Acute myelogenous leukemia Myeloproliferative disorder' Myelodysplastic syndrome Hodgkin's disease Non-Hodgkin's lymphoma Other Active malignancy. % Donor-recipient HLA-identity, % Total body irradiation, Oro GVHDt prophylaxis Methotrexate, % Cyclosporine, % Survival. days (range) • Includes chronic myelogenous leukemia (n

t Graft versus host disease.

1.134 28 (0.8-63) 57 81 232 314 324 43 38

348 30 (0.6-61) 59 85 86 89 78 15 21 47 12 62 58

(25 0Al) (26%) (22%) (4%) (6%) (14%) (3%)

(20%) (28 0Al) (29%) (4%) (3%) 144 (13%) 39 (3%) 48 72 81

84

81 71 505 (3-1.816)

75 80 78 (0-1,709)

= 395).

TABLE 2 ADJUSTED RELATIVE RISKS FOR RESPIRATORY FAILURE AFTER MARROW TRANSPLANTATION Factor

p Value

RR (95% GIl'

Donor-recipient HLA-non-identity Active malignancy Age, per 10 yr

< 0.001

1.70 (1.33-2.16)

< 0.001 < 0.001

1.56 (1.39-1.75) 1.19 (1.14-1.24)

• Relative Risk (95% confidence intervals) adjusted for the covariates total body irradiation, radiation dose. and primary

disease of Hodgkin's disease. These covariates were found'to be significantly associated with respiratory failure in Cox proportional hazards model in the absence of HLA·identity, malignant activity status, and age.

512

CRAWFORD AND PETERSEN

/~

NO VENTILATOR .'1134

VENTILATOR .'348 (2311)

I

EXTUBATED .'72 (2111)

l~IED

DISCHARGED

.'67

r----!

n-16

36 RECEIVED ADDITIONA.L

~ VENTILATION

DIED .'333 (2911) 188 (80"11' WITH RELAPSE OF

ALIVE AT 6 MONTHS

ALIVE AT 8 MONTHS

"-S01 (71%)

"-10 (3 ..)

DIED

:~ITH

organ systems at the time of ventilatory support. None of the long-term survivors of mechanical ventilation had a respiratory infection documented as a cause for respiratory failure. Follow-up information 1 yr after transplantation was available for nine of these 10 long-term survivors, and eight performed repeat pulmonary function tests (table 3). In general, testing was normal prior to transplant, except for a mild restrictive defect in one (table 3, patient 3) and decreased diffusing capacity in four (table 3, patients 3, 4, 8, and 9). At 1 yr after transplant, three patients had minimal respiratory complaints. Two noted cough and recurrent upper respiratory infections (table 3, patients 4 and 7). Both of these patients demonstrated airflow obstructive defects at 1 yr (reduced FEV l/FVC). One patient had mild dyspnea on climbing one flight of stairs (table 3, patient 5), and pulmonary function testing revealed moderate restrictive defect (TLC, 55% of predicted). The remaining patients, although asymptomatic, demonstrated reductions in singlebreath diffusing capacities. Arterial blood gas determinations were not routinely obtained at 1 yr reevaluation.

Fig. 1. Schematic of outcome of 1,482 patients undergoing marrow transplantation for treatment of malignancy between January 1986 and July 1990.Assisted ventilatory support was provided to 348 marrow recipients (23%) after transplantation. Patients requiring mechanical ventilation for less than 24 h postoperatively or after grand mal seizure (n = 34) were not included as receiving mechanical ventilation.

RELAPSE

OF MALIGNANCY

MALIGNANCY

not shown). Thus, total body irradiation or dose did not appear to be independent risk factors for respiratory failure.

Survivors of Mechanical Ventilation Mechanical ventilatory support was discontinued, and extubation was performed in 72 patients (21070) (figure 1). Fifty-seven (80%) of these patients died in hospital (or terminal care facility), and repeat ventilatory support was required by 36 (50%) of these patients. The interval between episodes of mechanical ventilation was from zero to 59 days (mean, 19 days). No patient requiring repeated mechanical ventilation survived to hospital discharge. Fifteen marrow recipients (4% of patients ventilated) were discharged to home (figure 1). Five of these patients died within the next 6 months. Three died after relapse of the malignancy, one died from acute respiratory failure, and one because of sepsis. Long-term survival (6 months after transplant) was achieved by 10 patients (3%) (figure 1 and table 3). Ventilatory support had been provided for an average of 6 days (range, 1to 9 days) for these patients. Most of these episodes occurred

within the first 3 wk after transplant (mean, 20 days; range, 11 to 48 days). Seven of these 10patients required ventilatory support for acute oxygenation failure. Four of these also required hemodialysis for acute renal failure. In addition, hepatic veno-occlusive disease was common. Four of the seven patients with hypoxemic respiratory failure had bilirubin levels above 3.0 mg/dl with the highest of 32. Most patients were neutropenic during the episode of respiratory failure. Only one of these seven patients with acute oxygenation failure had an absolute neutrophil count above I,OOO/mm 3 • The other three patients received ventilatory support primarily for hypercarbic ventilatory failure; severe oral mucositis with obtundation and airway obstruction (table 3, patient 7), obtundation with hypoventilation after an intracranial hemorrhage (table 3, patient 8), and prolonged ventilatory support after an open lung biopsy for idiopathic pneumonia (table 3, patient 6). All three of these patients had hepatic veno-occlusive disease with elevated bilirubin levels (3.6 to 5.9). In all, five of the long-term survivors had severe dysfunction of at least four

Discussion This study confirms the high prevalence of respiratory failure after marrow transplantation for malignancy and high mortality rate previously noted (4,5). In addition, risk factors noted in a previous study of 238 marrow recipients requiring mechanical ventilatory support at this center between 1980 and 1985 are supported by this analysis, which is the largest cohort of marrow recipients examined to date for pulmonary complications (4). The risk factors most strongly

TABLE 3 PULMONARY FUNCTION CHANGES AMONG SURVIVORS OF RESPIRATORY FAILURE Before Transplant Patient No.

Age

Day Onset'

Days Ventilated

23 20 11 20 11

73 77 93 84 82

(yr)

Sex

26 31 13 5 13 20 57 47 17

12 23 19 15

8 9 8 6 2 1 5 7 5 8

20

6

1 2 3 4 5 6 7 8 9 10

48

M F F F M M M F F M

Average

28

5.5

48

FEV,'FVC 68 73 86 91 83

88

%TLC

96 71 121 79 96 103 101 78

One Year after Transplant

%DLCO

AaP0 2

FEV,IFVC

%TLC

%DLCO

73 27 17

96

68

107 100 75 186 55 70 91

57

88

58 82 32 50 71

87

72

68

80

95

60

68

54 102 85 84

26 27

96 46 83 93

65 49 102

9

88 92

78

21

• Number of days aftar marrow transplant that mechanical ventilation was instnuted.

,AaP0 2

24

513

SURVIVAL FROM RESPIRATORY FAIWRE AFTER MARROW TRANSPLANTATION

and independently associated with the subsequent requirement of assisted mechanical ventilation were confirmed to be older patient age, active malignant disease at time of transplant, and receipt of a HLA-non-identical allogeneic donor marrow. The prevalence of mechanical ventilation after marrow transplantation changed little between studies. The percentage of patients surviving a first episode of mechanical ventilation was lower in the current study (21 versus 27070). This may represent an increased mean age of patients undergoing transplantation (29 versus 24 yr); however, long-term survival was similar (3 versus 5070). The present study was limited to patients with malignancy since the prevalence of respiratory failure after transplantation for aplastic anemia is low (4). We believe that mechanical ventilation is a valid surrogate for respiratory failure. It is not known what percentage of patients with respiratory failure either refused or were not offered ventilatory support. The prevalence of mechanical ventilation may underestimate the true prevalence of respiratory failure. However, the proportion of marrow recipients receiving ventilatory support has not changed since our previous study (4). Prior studies of intensive care for respiratory failure of patients with cancer (21-23), hematologic malignancies (24, 25), and marrow transplantation (5) had reported low survival rates. However, these studies involved relatively small numbers of patients. In addition, few patients requiring mechanical ventilation for longer than 6 days survived hospitalization. For example, Denardo and coworkers (5) reported no survivors among patients requiring> 4 days of mechanical ventilation (but noted that the 95% confidence intervals were zero and 12.5070) (5). The present study found a 3070 longterm survivor rate. Although this prevalence of survivors is similar to other smaller cohorts of patients with malignancy that have been reported, the large sample size permits a calculation of the 95070 confidence intervals of 2 to 6070. In distinction to prior smaller studies, eight of the 10 long-term survivors were ventilated for more than 4 days. The reasons for this difference is most likely related to the larger number of patients in our cohort. In addition, we excluded patients ventilated for brief periods after surgical procedures. Such patients were included in the study by Denardo and coworkers (5) and represented a proportion of

the reported survivors of intensive care after marrow transplantation. Multiple organ failures have been associated with death in the intensive care unit in several studies (21, 26, 27). Admittedly, the prognosis from respiratory failure after marrow transplantation is very poor; however, survival from multiple organ failures is not unprecedented. Several of the survivors noted in this study required hemodialysis, and half had as many as four organ dysfunctions. It is likely that these cases represent the extreme examples of favorable clinical outcome, but they do demonstrate the power of large cohort studies in assessing the accurate prevalence of outcomes. The results of this study can be viewed in various ways. The low incidence of long-term survivors can be taken to mean that the prognosis is uniformly grim (24). Such a conclusion may imply that medical intervention would be futile. Schneiderman and coworkers (28) have proposed that a quantitative approach to medical futility can be offered. They suggest that if experience demonstrates that the success is less than 1070 that medical intervention may be considered futile. On the basis of this study, the probability of survival from respiratory failure after marrow transplantation, even with multiple organ failure, appears to be at least 2%. Thus, medical futility may not be a valid argument for withholding mechanical ventilation from the marrow transplant recipient unless the patient (or appropriate surrogate) concurs. Another view of the data would be that long-term survival is possible. The decision regarding continued treatment should be made by the patient (or surrogate) on the basis of the probabilities and likely burdens imposed by the treatment. Given the relatively young age of many of the marrow transplant recipients and the prospects for long-term survival, optimism may be an appropriate view to take of the data. No patient in the present study who required ventilatory support for longer than 9 days was a long-term survivor. However, we have previously reported and have subsequently seen survivors ventilated beyond this time (4). Prolonged duration of ventilation likely represents a very poor prognosis, but survival is not unprecedented. It is possible that recurrent respiratory failure, or respiratory failure caused by pulmonary infection, may exemplify unprecedented survival and a truly futile situation. However, caution should

be exercised since the numbers of patients fulfilling these criteria are limited. In addition, effective treatments of significant transplant-related infections (such as cytomegalovirus pneumonia) continue to be developed. Our ability to predict on the basis of patient characteristics who will survive respiratory failure is limited (4). Given the difficulty in assessing medical futility in the marrow transplant recipient with respiratory failure, autonomous decisions by the patient should be followed (29). Advanced-care directives can be obtained prior to marrow transplantation from marrow recipients at risk for respiratory failure. The risk factors and the prevalence of outcomes presented here can be used in counseling before marrow transplantation. References I. Press OW, Schaller RI, Thomas ED. Bone marrow transplant complications. In: Toledo-Pereyra LH, ed. Complications of organ transplantation. New York: Marcel Dekker, 1987; 399-424. 2. Thomas ED. Marrow transplantation for malignant diseases (Karnofsky Memorial Lecture). J Clin Oncol 1983; 1:517-31. 3. Krowka MJ, Rosenow EC, Hoagland HC. Pulmonary complications of bone marrow transplantation. Chest 1985; 87:237-46. 4. Crawford SW, Schwartz DA, Petersen FB, Clark JO. Mechanical ventilation after marrow transplantation: risk factors and clinical outcome. Am Rev Respir Dis 1988; 137:682-7. 5. Denardo SJ, Oye RK, Bellamy PE. Efficacy of intensive care for bone marrow transplant patients with respiratory failure. Crit Care Med 1989; 17:4-6. 6. Thomas ED, Buckner CD, Storb R. Aplastic anemia treated by marrow transplantation. Lancet 1972; 1:284-9. 7. Thomas ED, Storb R. Techniques for human marrow grafting. Blood 1970; 36:507-15. 8. Thomas ED, Storb R, Clift RA, et al. Bonemarrow transplantation. N Engl J Med 1975; 292:832-43, 895-902. 9. Storb R, Thomas ED. Allogeneic bone-marrow transplantation. Immunol Rev 1983; 71:77-102. 10. Sullivan KM, Storb R. Allogeneic marrow transplantation. Cancer Invest 1984; 2:27-38. 11. Deeg HJ, Storb R, Thomas ED, et aL Cyclosporine as prophylaxis for graft-versus-host disease: a randomized study on patients undergoing marrow transplantation for nomlymphoblastic leukemia. Blood 1985; 65:1325-34. 12. Storb R, Deeg HJ, Thomas ED, et al. Marrow transplantation for chronic myelocytic leukemia: a controlled trial of cyclosporine versus methotrexate for prophylaxis of graft-versus-host disease. Blood 1985; 66:698-702. 13. Irle C, Deeg HJ, Buckner CD, et al. Marrow transplantation for leukemia following fractionated total body irradiation: a comparative trial of methotrexate and cyclosporine. Leuk Res 1985; 9:1255-61. 14. Petersen FB, Buckner CD, Clift RA, et al. Laminar air flow isolation and decontamination: a prospective randomized study of the effects of prophylactic systemic antibiotics in bone marrow transplant patients. Infection 1986; 14:115-21.

514

15. Hughes WT, Rivera GK, Schell MJ, Thornton D, Lott L. Successful intermittent chemoprophylaxis for pneumocystiscariniipneumonitis. N Engl J Med 1987; 316:1627-32. 16. Gardner RM. Snowbird workshop in standardization of spirometry. ATS statement. Am J Respir Dis 1979; 119:831-8. 17. Goldman HI, BecklakeMR. Respiratoryfunction tests. Normal values at median altitudes and the prediction of normal results. Am Rev Tuberc 1958; 79:457-67. 18. Crapo RO, Morris AH, Gardner RM. Refer-

ence spirometricvaluesusing techniques and equipment that meet ATS recommendations. Am Rev Respir Dis 1981; 123:659-64. 19. Crapo RO, Morris AH. Standardized single breath normal values for carbon monoxide diffusing capacity. Am Rev Respir Dis 1981; 123:185-9.

CRAWFORD AND PETERSEN

20. Gilbert R, Auchincloss JH Jr. The interpreta-

tion of the spirogram: how accurate is it for obstruction? Arch Intern Med 1985; 145:1635-9. 21. Hauser MJ, Tabak J, Baier H. Survival of patients with cancer in a medical intensive care unit. Arch Intern Med 1982; 142:527-9. 22. Snow RM, Miller WC, Rice DL, Ali MK. Respiratory failure in cancer patients. JAMA 1979;

25. PetersSG, MeadowsJA, GraceyDR. Outcome of respiratory failure in hematological malignancy. Chest 1988; 94:99-102. 26. Bartlett RH, Morris AH, Fairly HB, Hirsch R, O'Connor N, Pontoppidan H. A prospective study of acute respiratory failure. Chest 1986; 89:684-9. 27. Gillespie OJ, Marsh HMM, Divertie MB,

Balakrishnan PV.Outcome of lung cancer patients requiring mechanicalventilationfor pulmonary failure. JAMA 1986; 256:3364-6. 24. Schuster DP, Marion JM. Precedents for meaningful recovery during treatment in a medical intensive care unit: outcome in patients with hematological malignancy. Am J Med 1983; 75:

Meadows JA. Clinical outcome of respiratory failure in patients requiring prolonged (> 24 hours) mechanical ventilation. Chest 1986; 90:364-9. 28. Schneiderman LJ, Jecker NS, Jonsen AR. Medical futility: its meaning and clinical implications. Ann Intern Med 1990; 112:949-54. 29. Schneiderman LJ, Arras JD. Counseling patients to counsel physicians on future care in the event of patient incompetence. Ann Intern Med

402-8.

1985; 102:693-8.

241:2039-42. 23. Ewer MS, Ali MK, Atta MS, Morice RC,