Bone Marrow Transplantation, (1998) 22, 787–794  1998 Stockton Press All rights reserved 0268–3369/98 $12.00 http://www.stockton-press.co.uk/bmt

Outcomes following mechanical ventilation in children undergoing bone marrow transplantation AB Warwick1, AC Mertens2, X Ou Shu2, NKC Ramsay3 and JP Neglia4 1

Division of Pediatric Hematology/Oncology, Department of Pediatrics, Medical College of Wisconsin, Milwaukee, WI; 2Division of Pediatric Epidemiology and Clinical Research, 3Bone Marrow Transplant Program, Department of Pediatrics, and 4Division of Pediatric Hematology/Oncology, Department of Pediatrics, University of Minnesota, Minneapolis, MN, USA

Summary: Between 1976 and 1992, 869 patients ⬍19 years of age underwent BMT at the University of Minnesota for a variety of malignant and non-malignant disorders. One hundred and ninety-six required mechanical ventilation (MV) at some time from the start of pre-BMT cyto reduction through the first year following BMT. Reasons for MV included respiratory compromise, upper airway management and non-pulmonary indications for respiratory support. In multivariate models, underlying diagnosis, receipt of HLA-mismatched marrow and the presence of acute graft-versus-host disease (aGVHD) were independent predictors of the need for MV. Indication for MV, underlying diagnosis, and presence of aGVHD were independent predictors of successful extubation. Overall survival at 2 years was 14% among MV patients and 52% among non-MV patients. While the need for MV during BMT reduces the overall likelihood of survival, 40% of children who required MV were successfully extubated; 35% of these extubated patients were long-term survivors. This outcome is better than that reported for adult BMT patients requiring respiratory support, who show survival of ⬍5% at 6 months following BMT. Our data suggest extrapolation of outcome data from adult to pediatric patients is not appropriate and aggressive care of pediatric patients requiring respiratory support is not futile. Keywords: mechanical ventilation; marrow transplantation; children; pulmonary complications

Bone marrow transplantation (BMT), autologous and allogeneic, has become part of standard therapy for patients with leukemia, solid tumors, immunologic diseases, aplastic anemia and metabolic disorders correctable by infusion of normal hematopoietic stem cells.1,2 Much of the morbidity associated with BMT is due to complications arising from the transplant process including direct toxicities of the radiation and/or drugs used in preparative regimens, as well as infections and bleeding complications that result from Correspondence: Dr AB Warwick, Department of Pediatrics, Medical College of Wisconsin, 8701 Watertown Plank Road, Milwaukee, WI 53226, USA Received 9 March 1998; accepted 28 May 1998

the prolonged marrow suppression. In addition, graftversus-host disease (GVHD) can add to morbidity and mortality in recipients of allogeneic marrow.1 As preparative regimens for bone marrow transplantation become more aggressive, patients require more intensive support, which may include mechanical ventilation (MV). The use of MV has been associated with a poor outcome for bone marrow transplant recipients. In two retrospective studies, ⬍5% of all patients placed on ventilators during BMT were reported to survive 6 months after extubation;3,4 most of these patients were adults. One of the tenets of pediatric practice is that children are not small adults. Our clinical impression prior to this study suggested that children requiring MV during BMT had a better prognosis than that reported in the literature. An understanding of the risks and benefits associated with MV for respiratory failure or other indications is important for both pre-transplant counseling and informed choices during and after transplant by physicians, patients, and their families. We examined a 16year cohort of pediatric BMT patients to identify the characteristics of patients who required MV support and the factors associated with patient outcome after MV.

Materials and methods Subjects All patients less than 19 years of age at time of BMT at the University of Minnesota between January 1976 and December 1992 were included in the study (n = 896). Of these, 196 required MV at some time from the start of preBMT cytoreduction through the first year following BMT. As of July 1993, the minimum follow-up period from time of transplant was 6 months. Methods The University of Minnesota BMT database contains demographic information of each patient transplanted at the University Hospital as well as clinical data collected prospectively during and after transplant. This database was used to extract a complete dataset of eligible patients along with information about their conditioning regimen, date and type of BMT, status of HLA match, GVHD prophylaxis, original disease and remission status, current status and status date, and details of pulmonary problems during or after the trans-

Mechanical ventilation after pediatric BMT AB Warwick et al

788

plant. One investigator (ABW) reviewed the complete medical records of each patient who required MV and extracted into a separate dataset the reason for intubation, duration of MV, reason for extubation, and time after extubation to re-intubation, death or continued survival. We combined the information from both datasets for analysis and for comparison with the total pediatric BMT population during this interval. Marrow transplantation All parents, and the patients who were old enough, consented to participate in BMT protocols approved by the Human Subjects Committee at the University of Minnesota. Patients received autologous, syngeneic, or allogeneic (related or unrelated, HLA-matched or mismatched) marrow. Conditioning regimens varied by disease and type of BMT. Most patients received cyclophosphamide with either busulfan or total body irradiation (TBI). For GVHD prophylaxis patients received methotrexate and prednisone, with either cyclosporin A or anti-thymocyte globulin. After 1980 patients stayed in an Hepa-filtered isolation room; before that time they stayed in a conventional hospital room with positive-pressure airflow. For prophylaxis of Pneumocystis carinii infection, patients received either oral trimethoprim/sulfamethoxazole or aerosolized pentamidine from time of admission until at least 6 months after BMT. A regimen of broad-spectrum antibiotics was started with the first fever (T ⬎ 38.5°C), and continued until the absolute neutrophil count was greater than 500. In patients at risk for infection with cytomegalovirus, herpes simplex virus or varicella zoster virus, prophylaxis with acyclovir and/or immune globulin was started in 1985. Mechanical ventilation and survival MV was defined as the delivery of a mechanically generated tidal volume through an endotracheal tube. Intubation for surgical procedures was excluded if extubation was successful at the end of the operation. Patients were considered to be successfully extubated if they were alive without repeat MV support at least 1 day after extubation, to be short-term survivors if they were alive 30 days after extubation, and to be long-term survivors if they were alive 6 months after extubation. Statistics Patient demographic, disease and bone marrow transplantation variables were analyzed for association with mechanical ventilation and extubation by univariate analysis using standard parametric statistics. The indications for MV were analyzed as predictors of successful extubation using logistic regression. Cox regression models were employed to estimate the relative risk of requiring mechanical ventilation associated with studied variables and to allow censoring of deaths occurring during the first year post-BMT. Survival from extubation was calculated using the Kaplan– Meier method, and comparisons of the survival of patients requiring and not requiring mechanical ventilation was conducted using the log-rank test statistic. All tests of statistical

significance were two sided; P values of 0.05 or less were considered significant. In analyzing acute graft-versus-host disease (aGVHD) as a risk factor for MV or a predictor of successful extubation, the time of aGVHD onset with respect to the occurrence of MV was considered. For cases where aGVHD occurred after the initiation of MV, aGVHD was not considered as a risk factor for initiation of MV. Similarly, aGVHD occurring after extubation was not considered as a risk factor for successful extubation.

Results Patient population (Table 1) Between 1 January 1976 and 31 December 1992, 869 pediatric patients received a BMT for treatment of a non-malignant or malignant disease such as aplastic anemia, immune deficiency, metabolic disorder, acute and chronic leukemia, lymphoma or neuroblastoma. One hundred and ninety-six patients required MV at some time from the start of pre-BMT cytoreduction through the first year following BMT. Forty-five patients had two or more separate courses of mechanical ventilation, fourteen patients received at least three, and two patients underwent four. In total, 254 episodes of ventilatory support occurred in these 196 patients. The primary reasons for intubation and mechanical ventilation for the first episode of MV, as recorded by the responsible physician at the time of intubation, are listed in Table 2. The most frequently given reason for intubation was intrinsic respiratory distress (whether infectious or noninfectious was often not specified in the medical record at the time of intubation). Pneumonia, pneumonitis, acute respiratory distress syndrome, or other respiratory decompensation accounted for intubation in one-third of the patients. Other reasons for intubation included protection and control of the upper airway in patients with severe mucositis or seizures, cardiac arrest in patients with underlying sepsis, impending respiratory failure in patients with sepsis alone, extrinsic respiratory compromise from ascites or venoocclusive disease of the liver, prolonged intubation following non-pulmonary surgical procedures, pulmonary surgery and/or related procedures, eg broncho-alveolar lavage, and cardiac arrest following upper respiratory obstruction. Mechanical ventilation was initiated a median of 30 days post BMT; with a range from 12 days before BMT, for upper airway control in a child with Hurler’s syndrome, to 348 days following BMT, for abdominal distention and respiratory distress. The duration of MV ranged from 1 day, following cardiovascular arrest, to 102 days, for slow progressive respiratory failure, with a median of 8 days. Risk factors for MV (Table 1) Patients requiring MV did not differ from non-MV patients by age, sex or pre-BMT radiation schedule. In univariate analysis, patients undergoing BMT for immune deficiency, metabolic disease and neuroblastoma were more likely to require MV than those undergoing BMT for other indi-

Mechanical ventilation after pediatric BMT AB Warwick et al

Table 1

789

Patient characteristics and risk factors for mechanical ventilation Mechanical No mechanical ventilation (n = 196) ventilation (n = 673) (%) (%)

Age Mean s.e.m. Sex Male Female Disease Aplastic anemia Immune deficiency Metabolic disease Acute leukemia CML/MDS Lymphoma Neuroblastoma Other Donor Related Unrelated Autologous HLA Match Mismatch aGVHD None Grade 1 Grade 2 Grade 3 Grade 4 TBI None TLI TBI FTBI HFTBI a

Relative risk

95% Confidence interval

P value

7.52 0.4

8.39 0.21

107 (54.6) 89 (45.4)

392 (58.2) 281 (41.8)

1.0 1.1

— 0.8–1.5

61 33 43 375 55 41 46 19

(9.1) (4.9) (6.4) (55.7) (8.2) (6.1) (6.8) (2.8)

1.0a 3.1 1.8 0.7 0.7 0.4 1.5 0.4

— 1.8–5.4 0.9–3.3 0.4–1.3 0.3–1.5 0.1–1.2 0.8–2.9 0.1–0.8

111 (56.6) 43 (21.9) 42 (21.4)

356 (52.9) 62 (9.2) 255 (37.9)

1.0a 2.1 0.6

— 1.5–3.0 0.4–0.9

⬍0.01 ⬍0.01

134 (68.4) 62 (31.6)

602 (89.5) 71 (10.5)

1.0a 3.0

— 2.2–4.2

⬍0.01

a

— 0.7–2.2 0.6–1.6 1.1–2.6 1.9–5.2

⬍0.05 ⬍0.01

18 39 24 77 11 4 21 2

(9.2) (19.9) (12.2) (39.3) (5.6) (2.0) (10.1) (1.0)

125 13 19 21 18

(63.8) (6.6) (9.7) (10.7) (9.2)

503 38 71 44 17

(74.7) (5.6) (10.5) (6.5) (2.5)

1.0 1.2 1.0 1.7 3.1

56 11 59 49 21

(28.6) (5.6) (30.1) (25.0) (10.7)

145 46 190 198 94

(21.5) (6.8) (29.4) (29.4) (14.0)

1.0a 0.7 0.8 0.7 0.7

⬍0.01

— 0.3–1.3 0.6–1.2 0.5–1.1 0.4–1.1

Reference category.

Table 2

Reasons for mechanical ventilation (n = 196)

Cause for initiation of mechanical ventilation Primary pulmonary disease Respiratory, intrinsic Respiratory, compromise from extrinsic compression Post-op: non-pulmonary surgery Non-pulmonary causes Sepsis, cardiovascular collapse Arrest: due to sepsis, cardiovascular collapse, etc Post-op: pulmonary surgery Upper airway management Arrest: Airway problems Airway control

n

(%)

71 13

(36.2) (6.6)

7

(3.6)

24 25

(12.2) (12.8)

10

(5.1)

4 42

(2.0) (21.4)

marrow from an unrelated donor were twice as likely to require MV than a patient receiving marrow from a sibling. Patients receiving autologous grafts were statistically less likely to require MV when compared to the reference sibling group. The use of mismatched bone marrow, marrow from a related donor who was not a complete six antigen match for the recipient, was associated with a three-fold increased likelihood of MV. Among patients receiving allogeneic marrow transplants, those who developed grades 3– 4 aGVHD were more likely to require MV than patients with grades 0–2 aGVHD. In multivariate models, underlying diagnosis, HLA mismatched marrow, and aGVHD were all found to be independent significant predictors of the need for MV; marrow donor source was not. Predictors of successful extubation (Table 3)

cations when compared to our reference group, patients undergoing BMT for aplastic anemia. Of these diagnoses, only the increased association with immune deficiency reached statistical significance. Marrow source and degree of matching were statistically associated with the likelihood of MV. Patients receiving

Seventy-nine of the 196 MV patients were considered successfully extubated. In univariate analysis, the likelihood of successful extubation varied by underlying disease, reason for intubation, grade of aGVHD and radiation schedule; we found no relationship between duration of MV and successful extubation. Patients intubated following BMT for meta-

Mechanical ventilation after pediatric BMT AB Warwick et al

Table 3

Predictors of successful extubation

Sex Male Female Disease Aplastic anemia Immune deficiency Metabolic disorder Acute leukemia CML/MDS Lymphoma Neuroblastoma Other Donor Related Unrelated Autologous HLA Match Mismatch aGVHD None Grade 1 Grade 2 Grade 3 Grade 4 TBI None TLI TBI FTBI HFTBI Reasons for MV Respiratory, intrinsic Respiratory, extrinsic (restrictive) Post-op: pulmonary surgery Sepsis, CV collapse Arrest: sepsis, etc Post-op: non-pulmonary surgery Arrest: Airway problems Airway control

Successful extubation (n = 79) (%)

No successful extubation (n = 117) (%)

Relative risk

47 (59.5) 32 (40.5)

60 (51.3) 57 (48.7)

1.0 0.7

0.4–1.3

5 9 16 32 5 2 9 1

13 30 8 45 6 2 12 1

(11.1) (25.6) (6.8) (38.5) (5.1) (1.7) (10.3) (0.9)

1.0a 0.8 5.2 1.8 2.2 2.6 2.0 2.6

— 0.2–2.8 1.3–19.8 0.6–5.7 0.5–10.4 0.3–23.8 0.5–7.5 0.1–50.0

41 (51.9) 17 (21.5) 21 (26.6)

70 (59.8) 26 (22.2) 21 (17.9)

1.0a 1.1 1.7

— 0.5–2.3 0.8–3.5

56 (70.9) 23 (29.1)

78 (66.7) 39 (33.3)

1.0a 1.0

— 0.4–1.5

41 6 8 13 11

(51.9) (7.6) (10.1) (16.5) (13.9)

84 7 11 8 7

(71.8) (6.0) (9.4) (6.8) (6)

1.0a 1.8 1.5 3.3 3.2

— 0.6–5.6 0.6–4.0 1.3–8.7 1.2–8.9

29 2 17 21 10

(36.7) (2.5) (21.5) (26.6) (12.7)

27 9 42 28 11

(23.1) (7.7) (35.9) (23.9) (9.4)

1.0a 0.2 0.4 0.7 0.8

— 0.0–1.0 0.2–0.8 0.3–1.5 0.3–2.3

20 6 2 9 7 6 2 27

(25.3) (7.6) (2.5) (11.4) (8.9) (7.6) (2.5) (34.2)

51 7 8 15 18 1 2 15

(43.6) (6.0) (6.8) (12.8) (15.4) (0.9) (1.7) (12.8)

1.0a 2.2 0.6 1.5 1.0 15.3 2.6 4.6

(6.3) (11.4) (20.3) (40.5) (6.3) (2.5) (11.4) (1.3)

95% Confidence interval

P value

⬍0.05

⬍0.05 ⬍0.05

⬍0.05

— 0.7–7.3 0.1–3.3 0.6–4.1 0.4–2.7 1.7–135.2 0.3–19.4 2.0–10.4

⬍0.05 ⬍0.05

a

Reference category.

bolic disease were much more likely to be successfully extubated than those transplanted for aplastic anemia or immune deficiency. Successful extubation was also much more likely when MV was initiated for upper airway management or non-pulmonary causes than for primary pulmonary disease. In our series, MV patients with grades 3– 4 aGVHD were successfully extubated more often than were MV patients with lower grades of aGVHD. MV patients who received single fraction TBI were successfully extubated less often than were those receiving fractionated or hyper-fractionated TBI. In multivariate analysis the indication for MV, underlying diagnosis and presence of aGVHD were all independent predictors of successful extubation. Radiation schedule was not significant in the multivariate model. Survival after mechanical ventilation/extubation (Figure 1) Twenty-nine of the 196 patients who required MV were alive at the time of analysis. One hundred and seventeen

100

75

Percent

790

Not intubated BMT patients 50

Successfully extubated patients

25

All intubated patients 0 0

2

4

6

8

10

Time (years) Figure 1 Survival from date of marrow transplantation.

12

14

Mechanical ventilation after pediatric BMT AB Warwick et al

died either on the ventilator or within 24 h of extubation; 79 were considered successfully extubated. Fifty-five of the 79 patients who were successfully extubated survived at least 30 days, and 33 survived more than 6 months after extubation. Although 151 of the 196 patients were placed on the ventilator only once, 45 required two or more courses of mechanical pulmonary support. Survival or time to death after extubation was not significantly different between patients who had a single episode of MV and those who were placed on the ventilator more than once (P = 0.42). In univariate analysis, survival after MV depended only on underlying diagnosis. Patients with aplastic anemia or immune deficiency did poorly with 1 year survivals by Kaplan–Meier estimation of 5.6 and 7.7%, respectively, while patients with metabolic diseases had a 1 year survival of 40.2% in comparison to the total MV population’s 1 year survival of 17.6%. While fewer patients survived who were intubated for intrinsic respiratory disorders (eight of 71) or an arrest related to sepsis and circulatory collapse (one of 25) than did those who were placed on the ventilator for airway control (11 of 42) or respiratory support following non-pulmonary surgery (two of seven), this difference was not statistically significant. Survival was not significantly related to timing or duration of ventilation, pretransplant conditioning regimen or source and match of donor bone marrow. There was no difference in survival among MV patients with or without aGVHD. Multivariate analysis yielded no factors predicting survival. Kaplan–Meier estimates of survivorship from the time of BMT for patients requiring MV were 26.7% at 6 months and 17.6% at 1 year. Survival at 2 years following BMT was 14.2% among all MV patients, 34.9% among successfully extubated MV patients, and 51.6% among non-MV patients. Cause of death The most frequent causes of death (Table 4) in these patients were respiratory failure or multi-system organ failure. Several of these patients also suffered from a serious viral or fungal infection that was not the direct cause of Table 4

Primary cause of death for mechanically ventilated patients

Cause of death

Respiratory failure Multi-system failure Sepsis, CV collapse Cardiopulmonary arrest CNS event Relapse Hepatic failure No information Total X2 P value = 0.0016.

Total n (%) Successfully extubated patients

65 (38.9) 49 (29.3) 14 (8.4) 19 (11.4) 6 (3.6) 5 (3.0) 1 (0.6) 8 (4.8) 167 (100.0)

n

(%)

Not successfully extubated patients n (%)

17 13 3 3 3 4 0 7

(34) (26) (6) (6) (6) (8) (0) (14)

48 (41) 36 (30) 11 (9) 16 (14) 3 (3) 1 (1) 1 (1) 1 (1)

50 (100.0)

117 (100.0)

death. The cause of death was not always related to the reason for intubation. Discussion Previously reported risk factors for MV include older age (⬎21 years), active disease at time of transplant and receipt of HLA-mismatched marrow.3–8 In our analysis, the underlying diagnosis, receipt of HLA-mismatched marrow and stage of aGVHD were independent significant predictors for MV by multivariate analysis. No association was found between patient age or disease status and the need for MV. Diagnoses associated with an increased risk of requiring MV were immune deficiency, metabolic disorder and neuroblastoma. Patients with an underlying immune deficiency frequently come to BMT with pre-existing infections, eg Pneumocystis carinii and cytomegalovirus, predisposing them to pulmonary problems which may require intubation. Children with a metabolic disorder such as Hurler’s syndrome are often intubated for upper airway control during the conditioning phase of BMT. Patients with neuroblastoma may be at increased risk because of very aggressive chemotherapy before selection for BMT. Crawford et al3 found an increased incidence of MV among BMT patients of all ages transplanted for an underlying malignancy, especially those with active disease at time of BMT, when compared to those transplanted for non-malignant disorders. In a later study,4 they confirmed this increased risk for patients with an active malignancy and further suggested that patients with Hodgkin’s disease had an increased risk for MV. Noble did not find an association between underlying diagnosis and MV; others did not comment on the possibility.5,7–12 We found no association between remission status of the underlying disease and MV. In this series, patients receiving marrow from an HLAmismatched donor had a significantly greater risk for MV than those receiving marrow from any other source. This association has been previously reported.3,4,12 Patients with grades 3–4 aGVHD were also significantly more likely to require MV than patients with grades 0–2 aGVHD. Other investigators have reported an increase in pulmonary complications, including MV, with higher grades of aGVHD.7,10–12 Suggested mechanisms include an increased risk for CMV pneumonia12,13 and progressive airway obstruction from either acute bronchitis or bronchospasm that may lead to lower respiratory tract disease.10 Once patients are placed on MV, questions arise such as: ‘Will we be able to extubate the patient?’ ‘When will we be able to extubate the patient?’ And, ‘Is it worth continuing?’ Thus, it is important to be able to identify and define patient characteristics that will serve as predictors for successful extubation. In our series, successful extubation varied in both univariate and multivariate analyses according to the reason for intubation and the patient’s underlying disease. Multivariate analysis added the presence of aGVHD, specifically grades 3–4, as another independent predictor for successful extubation. The mechanism underlying the seemingly contradictory associations of aGVHD with need for MV and successful extubation is unclear. In general, others have found the

791

Mechanical ventilation after pediatric BMT AB Warwick et al

792

Table 5

Data summary from the various reports Incidence

Current series Todd7 Cholowski14 Crawford4 Nichols11 3

Crawford Faber-Langendoen6 Wagner19 Rubenfeld18 Denardo5 Hollmig20 Martin15 Paz21 Torrecilla8

Successful extubation

Long-term survival

n

(%)

n

%

n

%

196/869 54/285 29/256 348/1134

(22.6) (18.9) (11.3) (30.6)

79/196 6/54 — —

(40.3) (11.1) — —

33/196 5/54 3/29 10/348

(16.8) (9.3) (10.3) (2.9)

23/318

(7.2)

—

—

2/23

(8.7)

232/1089 191/653 25/134 865/3635 44/158 10/85

(21.3) (29.0) (18.7) (23.8) (27.8) (12.0)

63/232 30/191 5/25 — 4/44 4/10

(27.2) (15.7) (20.0) — (9.1) (40.0)

16/232 6/191 1/25 30/865 2/44 3/10

24/350 28/229 16/57

(6.9) (12.2) (28.1)

— — 1/16

— — (6.3)

2/24 1/28 1/16

Age (years)

Duration (months)

Range

Mean

⬍19 ⬍19 26–53 0.7–63

8.2 10.25 43 29

to discharge from BMT

⬍19

—

(6.9) (3.1) (4.0) (3.5) (4.5) (30.0)

100 days after MV 6 months after MV 6 months after MV 12 months after MV 12 months after MV 2 years after MV

— ⬎18 19–55 0.8–58 ⬎18 ⬎18

— 33 38 26.4a 30 36

(8.3) (3.6) (6.3)

not given not given not given

1–67 ⬎18 7–44

28 34.4 28

6 6 5 6

months months months months

after after after after

BMT BMT BMT BMT

a

Median age.

presence of aGVHD to be associated with an increased rate of intubation and a poor prognosis.4,7,11 High doses of steroids are considered to be helpful in treating the pulmonary inflammation that occurs as a consequence of pulmonary hemorrhage.16,17 The high doses of steroids and other immune suppressive agents currently used for aGVHD could also reduce pulmonary inflammation occurring from other disorders which may have resulted in intubation, and so permit extubation. This hypothesis will need study. Patients who were intubated for primary pulmonary disease were less likely to be extubated than patients who required respiratory support for non-pulmonary indications. Our rate of successful extubation overall (40.3%), and even after intubation for pulmonary causes (30%), is higher than that reported elsewhere: 27% overall from Crawford’s initial report of 1089 adult and pediatric BMT patients,3 16% overall from Faber-Langendoen et al’s report of 653 adult BMT patients,6 and 11% overall from Todd et al’s report of 285 pediatric patients7 (Table 5). An explanation for our better rate of extubation may be that our study included only pediatric patients. The patient mix may also be important, for Todd’s series did not include patients with immune deficiencies or metabolic disease. There are no comments about an association between diagnosis and successful extubation in previously published studies specifically analyzing pediatric patients,7,11 adults,5,6,14,21 or both adults and children.3,4,8,15 In these studies, patient diagnosis was classified mainly as either ‘malignant’ or ‘non-malignant’. In our analysis, both the highest and lowest rates of extubation were found in patients with non-malignant diagnoses: metabolic disorders and immune deficiency, respectively. No other disease categories, either malignant or non-malignant, were associated with either an increased or decreased likelihood of successful extubation. This association might have been noted in other series if the patient diagnosis groups had been more specific. Several investigators4,5,6,8,11 have suggested the duration

of ventilation to be inversely proportional to the likelihood of successful extubation. Torrecilla8 and Crawford and Petersen4 described respiratory support to be futile if continued for more than 7 or 9 days, respectively. However, Afessa9 and Crawford’s earlier report3 showed patient outcomes to be affected by the institution but not the duration of mechanical ventilation. We found no relationship between the duration of mechanical ventilation and successful extubation. A possible explanation for this observation is that our patient population is pediatric, encompasses children undergoing bone marrow transplantation for a variety of diseases, and includes those intubated for any reason. Many predictors for patient long-term survival have been suggested including: male gender,11 the absence of multisystem organ failure,7,8,11,18 freedom from primary pulmonary parenchymal disease5,8,22 or infection,20,22 timing of intubation,6,18,20 duration of intubation,5,6,8,11 the absence of aGVHD,7,11 and age at time of BMT in adults.6,18,19 The precise predictive value of these factors is unclear and limited. We found none of these to be significant predictors for survival in our patient population. Follow-up study in pediatric patients is warranted, especially of primary lung disease and timing of intubation: two valuable predictors of survival for adult BMT patients. Other proposed predictors include calcium,20 leukocyte20 and serum albumin19 levels and the Acute Physiology and Chronic Health Evaluation (APACHE) scores at time of intubation.18–20 To evaluate individual blood test results was beyond the scope of our investigation. During the time period we studied, APACHE scoring was not done. Using multivariate models we found no significant predictor of survival, although in the univariate analysis underlying disease was statistically significant. Some investigators found, as we did, no independent predictors for survival,3,9 while other reports did not specifically address the issue.4,10,12,14 The overall survival in our patient population is much higher than that reported elsewhere.3–8,14,15 In our series, 79 of 196 MV patients (40.3%) were successfully extubated

Mechanical ventilation after pediatric BMT AB Warwick et al

and 33 of the 196 (16.8%) survived at least 6 months after BMT. Thus, 33 of the 79 successfully extubated MV patients (41.8%) were alive 6 months post-BMT; this absolute survival rate approaches that of the 673 BMT patients not requiring MV (48.6%). In the two reports of pediatric BMT patients absolute survival rates of 8.7% (two of 23) are cited among MV patients followed to time of discharge from hospital following BMT11 and 9.3% (five of 54) of patients followed for 6 months post-BMT.7 There are many reports of an association between the need for MV and a poor outcome in BMT patients.3–6,8,15,18,21 Crawford et al3,4 reported survivals, by Kaplan–Meier estimation, of 6.9% at 100 days after extubation3 and 3% at 6 months after BMT among MV patients of all ages.4 FaberLangendoen et al6 reviewed the experiences of 191 adult BMT patients on MV: by Kaplan–Meier estimates, 15% were able to be extubated but fewer than 4% survived 6 months from extubation. Paz et al21 reported the outcomes of 28 adult patients requiring intensive care following BMT: only one (3.7%) survived to discharge from that unit. Rubenfeld and Crawford found survival at 6 months and 1 year following MV to be 4 and 3.5%, respectively, in patients of all ages.18 Absolute rates of survival have been reported that range from 2.9% (10 of 348)4 to 10.3% (three of 29)14 at 6 months following BMT. Thus, we find survival with MV during BMT to be better in children than in adults, and that the chances of survival, once a patient is successfully extubated, were better than previously reported. In view of concerns about medical costs and health care reform, some have questioned the usefulness of MV in BMT patients but have reached no conclusions.5,7–9 FaberLangendoen et al6 proposed, for certain subsets of adult BMT patients at higher risk for pulmonary complications, that MV be considered only in conjunction with welldesigned clinical research trials. We have shown extrapolation from adult data to pediatric patients is inaccurate. Our experience, and that of Nichols11 and Todd,7 indicates that aggressive care including MV is not futile. Nevertheless, care should be used in interpreting the data in our study for several reasons. Separation of diagnosis from treatment and/or associated conditions is difficult. This is particularly important in the pediatric BMT population, which tends to be more heterogeneous than the adult BMT population. We found underlying diagnosis to be significant both as a risk factor for intubation and as a predictor for successful extubation. Diagnosis may serve as a surrogate indicator of several measured and unmeasured characteristics including pre-BMT therapy and complications as well as unique complications that may arise during BMT. Data regarding prior therapy and complications were not available and thus not adjusted for in these analyses. Much has been written on MV and the presence of infectious intrinsic pulmonary disease.8,10,13,20,21,23 It is not always possible at the time of intubation to be certain if respiratory distress has an infectious (and if so, what type) or non-infectious cause. There were insufficient data to analyze the contribution of infectious vs non-infectious causes to the likelihood of successful extubation and ultimate survival. Although we found no association between MV and

cause of death, we did not consider other covariates such as renal failure or hepatic veno-occlusive disease, which may have affected successful extubation or survival. We focused our attention on risk factors for MV and predictors of successful extubation that would be based on clinical information available at time of intubation, in order to define problems clinicians and families face in making decisions about whether or not to intubate a child. Thus, we cannot comment well on the probable outcomes of patients who develop such additional complications. These covariates, shown to be important predictors for survival in adults,7,8,11,18 deserve further study in the pediatric population. Differences in criteria for MV, either over the years at our institution or between our and other institutions, could have accounted for the better survival reported here for children. If less ill patients were intubated or patients more likely to die were not, then the chances of successful extubation and survival would increase. However, since the incidence of MV in this study is similar to that reported elsewhere, it is unlikely that our criteria (the need to maintain an airway and treatment of impending respiratory failure or cardiopulmonary arrest) differ substantially from that of other institutions. In addition, the incidence of MV, successful extubation and long-term survival did not change significantly throughout the years of this study. Rubenfeld and Crawford, in an analysis of the Seattle BMT data, proposed guidelines for the institution of MV and/or prolonged life support18 that could appropriately be applied to any medical procedure. However, the guidelines were developed from the analysis of data on predominantly adult patients. We have shown that our pediatric population differs significantly in rates of successful extubation and long-term survival from those reported for adults. Consequently, guidelines for pediatric patients should reflect this difference. This report was intentionally descriptive in nature, to show there is a difference in outcomes between children and adults. Further studies are needed to investigate possible correlations between APACHE scores (or other surrogate measurements of patient status), the presence or absence of multi-system organ failure and/or hepatic veno-occlusive disease, and specific infections with patient outcomes. In summary, while the need for MV decreases the probability for survival, pediatric MV patients are more likely to be successfully extubated and, once extubated, are more likely to survive than are adult MV patients. These data, the risk factors and other data presented in this study can be used in counseling patients and their families before and during transplant regarding the risks and benefits associated with MV for respiratory failure or other indications.

Acknowledgements This work was supported by grants CA21737 and the Children’s Cancer Research Fund of the University of Minnesota.

793

Mechanical ventilation after pediatric BMT AB Warwick et al

794

References 13 1 Ramsay NKC. Bone marrow transplantation in pediatric oncology. In: Pizzo PA, Poplack DG (eds). Principles and Practice of Pediatric Oncology, 2nd edn. JP Lippincott: Philadelphia, 1993, pp 315–334. 2 Barrett J, McCarthy D. Bone marrow transplantation for genetic disorders (review). Blood Rev 1990; 4: 116–131. 3 Crawford SW, Schwartz DA, Petersen FB, Clark JG. Mechanical ventilation after marrow transplantation: risk factors and clinical outcome. Am Rev Respir Dis 1988; 137: 682–687. 4 Crawford SW, Petersen FB. Long-term survival from respiratory failure after marrow transplantation for malignancy. Am Rev Resp Dis 1992; 145: 510–514. 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 Faber-Langendoen K, Caplan AL, McGlave PB. Survival of adult bone marrow transplant patients receiving mechanical ventilation: a case for restricted use. Bone Marrow Transplant 1993; 12: 501–507. 7 Todd K, Wiley F, Landaw E et al. Survival outcome among 54 intubated pediatric bone marrow transplant patients. Crit Care Med 1994; 22: 171–176. 8 Torrecilla C, Corte´s JL, Chamorro C et al. Prognostic assessment of the acute complications of bone marrow transplantation requiring intensive therapy. Int Care Med 1988; 14: 393–398. 9 Afessa B, Tefferi A, Hoagland HC et al. Outcome of recipients of bone marrow transplants who require intensive-care unit support. Mayo Clin Proc 1992; 67: 117–122. 10 Krowka MJ, Rosenow EC, Hoagland HC. Pulmonary complications of bone marrow transplantation. Chest 1985; 87: 237–246. 11 Nichols DG, Walker LK, Wingard JR et al. Predictors of acute respiratory failure after bone marrow transplantation in children. Crit Care Med 1994; 22: 1485–1491. 12 Noble PW. The pulmonary complications of bone marrow

14

15

16 17

18

19

20 21 22 23

transplantation in adults (Medical Staff Conference). West J Med 1989; 150: 443–449. Sommer SE, Emanuel D, Groeger JS, Carlon GC. Successful management of CMV pneumonia in a mechanically ventilated patient. Chest 1991; 100: 856–858. Cholowski AM, Alberts WM, Zorsky PE, Elfenbein GJ. Outcome of patients requiring mechanical ventilation following autologous bone marrow transplantation. Am J Respir Crit Care Med 1994; 149 (Suppl. 1): A489. Martin C, Maraninchi D, Saux P et al. Need for mechanical ventilation in the treatment of acute infections complicating bone marrow transplantation (letter). Crit Care Med 1990; 18: 120–121. Chao NJ, Duncan SR, Long GD et al. Corticosteroid therapy for diffuse alveolar hemorrhage in autologous bone marrow transplant recipients. Ann Intern Med 1991; 114: 145–146. Metcalf JP, Rennard SI, Reed EC et al. Corticosteroids as adjunctive therapy for diffuse alveolar hemorrhage associated with bone marrow transplantation. Am J Med 1994; 96: 327–334. Rubenfeld GD, Crawford SW. Withdrawing life support from mechanically ventilated recipients of bone marrow transplants: a case for evidence-based guidelines. Ann Intern Med 1996; 125: 625–633. Wagner A, Staudinger T, Kofler J et al. Ergebnisse der intensivmedizinischen betreung von patienten nach knochenmarktransplantation. Wien Klin Wochenschr 1996; 108: 677– 682. Hollmig KA, Soehngen D, Leschke M et al. Long-term survival of recipients of allogeneic bone-marrow transplantation after mechanical ventilation. Eur J Med Res 1997; 2: 62–66. Paz HL, Crilley P, Weinar M, Brodsky I. Outcome of patients requiring medical ICU admission following bone marrow transplantation. Chest 1993; 104: 527–531. Randle CJ, Frankel LR, Amylon MD. Identifying early predictors of mortality in pediatric patients with acute leukemia and pneumonia. Chest 1996; 109: 457–461. Dunagan DP, Baker AM, Hurd DD, Haponik EF. Bronchoscopic evaluation of pulmonary infiltrates following bone marrow transplantation. Chest 1997; 111: 135–141.