Received: 11 January 2002Accepted: 4 October 2002Published online: 26 November 2002© Springer-Verlag 2002

Abstract Objective: To determineprognostic factors in child recipientsof hematopoietic stem cell transplan-tation from blood or bone marrow(BMT) requiring critical care. Design: Retrospective study of a cohort of patients. Setting: PediatricIntensive Care Unit (PICU) in a uni-versity tertiary care center. Patientsand participants: Child recipients ofBMT requiring PICU admission.Measurements and results: Of the 151children receiving transplants in ourinstitution, 44 (29.1%) had 49 ad-missions to the PICU. Mechanicalventilation (MV) was required in 34patients (69.4% of all admissions).Overall mortality was 31/44 (70.4%).Mortality in patients requiring MVand not requiring MV was 26/34(76.5%) and 5/10 (50%), respective-ly. The following variables were sig-nificantly associated with mortalityin the univariate analysis: male gen-der (P=0.02), older age (P=0.03),acute graft versus host disease(aGVHD) grades III or IV (P=0.01),severe hemorrhagic cystitis (P=0.01),

the diagnosis of lung injury (P=0.04),the need for MV (P=0.03) or for re-nal replacement therapy (P=0.02),the presence of respiratory (P=0.003),cardiovascular (P=0.009) or gastro-intestinal (P=0.01) failures, and thefailure of ≥3 organs (P=0.01). In themultivariate analysis, the presence ofaGVHD grades III or IV, male gen-der, severe hemorrhagic cystitis, andthe failure of ≥3 organs were foundto be independent predictors of mor-tality. Conclusions: The need for in-tensive care is common among childrecipients of a BMT. These patientshave a high mortality rate but somecomplications are reversible withcritical care support. Certain clinicalparameters are useful to establish arealistic prognosis and to optimizethe use of the available resources.

Keywords Bone marrow transplantation · Life-threateningcomplications · Intensive care · Pediatrics · Prognosis · Multiple organ failure

Intensive Care Med (2003) 29:91–96DOI 10.1007/s00134-002-1549-2 N E O N ATA L A N D P E D I AT R I C I N T E N S I V E C A R E

Adelaida LamasEnrique OtheoPurificación RosJosé Luis VázquezMaría Soledad MaldonadoArturo MuñozIsabel Martos

Prognosis of child recipients of hematopoietic stem cell transplantation requiring intensive care

Introduction

Hematopoietic stem cell transplantation from blood orbone marrow (BMT) is a widely used therapy for certainhigh-risk hematologic and solid malignancies as well asfor non-malignant diseases such as severe aplastic ane-mia, thalassemia, sickle cell disease, congenital immuno-deficiencies, certain inborn errors of metabolism, and os-teopetrosis [1]. Since many of those indications occur in

the pediatric age group, a significant number of BMT pa-tients are children [2].

BMT can result in serious complications includingthose derived from severe and prolonged myelosuppres-sion, toxicity of conditioning regimen, and graft vs hostdisease (GVHD) in allogenic grafts. Opportunistic infec-tions, septic shock, mucositis, severe bleeding, convul-sive disorders, massive intracranial hemorrhage, adversereactions to drugs, hepatic or renal failures, and multiple

A. Lamas · E. Otheo (✉) · P. Ros J.L. Vázquez · I. MartosPediatric Intensive Care Unit, Department of Pediatrics, Hospital Ramón y Cajal, 28034 Madrid, Spaine-mail: eotheodetejada@hrc.insalud.esTel.: +34-91-3231857Fax:.: +34-91-3368417

M.S. Maldonado · A. MuñozPediatric Oncology Unit, Department of Pediatrics, Hospital Ramón y Cajal, 28034 Madrid, Spain

organ failure can occur [3]. Pulmonary complications arecommon and include interstitial pneumonitis, infectiouspneumonia, cardiogenic and non-cardiogenic pulmonaryedema, bronchiolitis obliterans, and alveolar hemorrhagesyndrome [4]. It has been hypothesized that severalproblems associated with BMT could be similar to reper-fusion injury or septic shock, conditions in which endo-thelial injury and the subsequent generalized capillaryleak syndrome lead to multiple organ failure [5].

It has been estimated that 11–40% of BMT recipients,adults or children, require critical care because of the se-verity of the complications developing after BMT [4, 6,7, 8, 9, 10, 11]. Given the high mortality rate of these pa-tients, particularly those receiving MV, and the largeamount of resources they need, the determination ofmortality risk factors is of great clinical relevance. Dif-ferent studies have analysed variables predicting out-come, in order to avoid futile care in cases with a poorprognosis and no chance for survival [3, 4, 6, 7, 8, 9, 10,11]. The purpose of this study is to analyze the outcomeof our cohort of pediatric patients who required intensivecare after a BMT, and to identify prognostic variableswhich might predict outcome.

Patients and methods

The study cohort consisted of patients less than 18 years of ageadmitted to our PICU from August 1991 to February 2000 becauseof acute complications developing after a BMT. These patients re-ceived an allogeneic or autologous BMT in Hospital Ramón y Cajal, a tertiary care center for BMT in Madrid, Spain.

There were no established criteria for PICU admission. Illnessseverity was assessed following the CCS score [12]. Patients whohad received a BMT in the past and had been admitted to thePICU because of an unrelated condition were excluded from theanalysis. In addition, BMT recipients admitted to the PICU forpostoperative care or for diagnostic procedures were not included.Several patients were admitted to the PICU more than once. Suc-cessive admissions were only considered as separate events if theindications for the different admissions were different and the timeelapsed between the previous discharge and the new admissionwas at least 7 days.

The charts of these admissions were reviewed and the follow-ing demographic variables were registered: age, gender, underly-ing disease, reason for admission, type of transplant (allogenicfrom sibling-matched, sibling mismatched, haplo-identical or un-related donors or autologous), conditioning regimen (chemothera-py alone or chemotherapy and total body irradiation), and daysfrom BMT to PICU admission.

To evaluate illness severity, the degree of intensive care sup-port and the complications developed during the PICU stay, thefollowing variables were analysed: length of PICU stay, require-ment for MV, requirement of renal replacement therapy, days onMV, presence of neutropenia, type and number of organs failing,and the following diagnoses: lung injury (LI), acute respiratorydistress syndrome (ARDS), acute GVHD (aGVHD) grades III orIV, severe hemorrhagic cystitis, and infection.

LI was diagnosed when a fraction of inspired oxygen higherthan 0.6 or positive end-expiratory pressure greater than 5 cmH2Owere required after the first 24 h of MV, according to Rubenfeld etal. [3]. ARDS was diagnosed as per the American-European Con-sensus Conference on ARDS [13]. To evaluate the severity of

aGVHD, the grading system of Glucksberg et al. was used [14].Severe hemorrhagic cystitis was diagnosed when catheterisationand bladder irrigation were required for treatment or prevention ofobstructive uropathy. An infection was diagnosed based on thepresence of suggestive clinical signs and symptoms and/or a posi-tive culture from blood, urine or bronchoalveolar lavage fluid, oron diagnostic histopathologic findings on biopsy or necropsy.Neutropenia was defined as a total neutrophil count <0.5×109/lduring stay in the PICU. Organ failure was assessed in 7 organs asestablished by Wilkinson et al. [15].

Survivors were defined as being alive 30 days after PICU dis-charge. Long-term survivors were defined as those who were aliveon day 180 after PICU discharge.

Statistical analysis

Statistical analysis was performed using statistical software pack-age SPSS for Windows (release 10.0.7) and EPI Info 6.04 (Centersfor Disease Control, Atlanta, Ga., USA, 1994). Qualitative vari-ables were compared by the Chi square test with Yates’ correction.Homogeneity in quantitative variables was analyzed using Bartlett’s test. The ANOVA test was used in the comparison ofquantitative variables with homogeneity, and the Mann Whitneytest or two-sample Wilcoxon tests (Kruskal-Wallis for two groups)in the comparison of non-homogeneous quantitative variables.Multivariate analysis was made with stepwise forward logistic re-gression. The accepted level of statistical significance was 5%.

Results

During the study period, 151 children received a BMT inour center. Types of transplant were: 82 autologous, 1syngenic, and 68 allogenic (42 from an HLA identicalsibling donor, 2 from an haplo-identical donor, 1 from anHLA sibling mismatched donor, and 23 from unrelateddonors). Among them, 44 patients (29.1%) required 49admissions to the PICU. The characteristics of the pa-tients who required admission are outlined in Table 1.The most frequent reasons for PICU admission were pul-monary complications, hemodynamic failure, and neuro-logic disease. Fourteen patients (28.6% of admissions)were in CCS III criteria on admission and 35 (71.4% ofadmissions) were in CCS IV.

Outcome and supportive measures are shown in Table 2. The number of organ system failures rangedfrom 0 to 7 (mean 3.5, median 3), and in 77.6% of ad-missions ≥3 organs failing were diagnosed. The hemato-logical, respiratory, and cardiovascular organ systemswere the most frequently affected.

Ventilated patients

MV was required in 34 of the 49 admissions (69.4%).Patients on MV made up 22.5% of all pediatric BMT re-cipients. The indications for MV were: pulmonary tumorinfiltration in 1 patient (2.9%), pulmonary edema in 8(23.5%), pulmonary hemorrhage in 4 (11.8%), alveolarpneumonia in 9 (26.5%), interstitial pneumonia in 5

92

(14.7%), obliterans bronchiolitis in 1 (2.9%), airway ob-struction in 1 (2.9%), hemodynamic instability in 4(11.8%), and coma in 1 (2.9%). Thus, the indications forMV were pulmonary in 28 patients and nonpulmonary in6. LI and ARDS were diagnosed in 17 and 7 patients, re-spectively. The duration of MV ranged from 1 to 41 days(median 5 days).

Survival and predictive factors

The mortality rate of this cohort was 31/44 (70.4%).Eighteen of 49 admissions (36.7%) resulted in PICU dis-charge. The mortality rate of patients receiving and notreceiving MV was 26/34 (76.5%) and 5/10 (50%), re-spectively (P=0.003). The mortality rate was significant-ly higher (P<0.001) in those patients with pulmonary in-dications for MV as compared to those with non-pulmo-nary indications [23/28 (82.1%), and 3/6 (50%) respec-tively]. For the 44 patients, 6 (13.6%) were long-term

survivors with a median survival time after PICU dis-charge of 56 months (range 19–115 months).

The underlying diagnosis, type of donor, conditioningregimens, days from BMT to admission to the PICU,length of PICU stay, the diagnosis of ARDS, infection,neutropenia or isolated renal, hepatic, hematologic orneurologic failures were not significantly related to mor-tality (Tables 3 and 4). Mortality rate according to timefrom BMT to PICU admission (0–30 days: 55.5%;31–99 days: 85.7%; ≥100 days: 50%) was higher for pa-tients admitted between days 31 and 99 after BMT, butdifferences did not reach statistical significance.

The diagnosis of aGVHD grades III or IV, severehemorrhagic cystitis, LI, respiratory, cardiovascular orgastrointestinal failures, older ages, the need for MV orrenal replacement therapy, the failure of ≥3 organ sys-tems, and male gender were all significantly associatedwith mortality in univariate analysis (Tables 3 and 4,Fig. 1).

In multivariate analysis, the presence of aGVHDgrades III or IV, male gender, severe hemorrhagic cysti-tis, and the failure of ≥3 organ systems were all found tobe independent and significant predictors of mortality(P<0.05).

93

Table 1 Clinical characteristics of 49 admissions to PICU afterBMT. (BMT hematopoietic stem cell transplantation from blood orbone marrow, PICU pediatric intensive care unit)

Number of patients 44Number of admissions 49Sex (male/female) 21/28Median age (years) 9 (range 1–18)

DiagnosisHematologic neoplasia 32Solid tumor 6Non malignant hematologic disease 7Non-hematological congenital disease 4

Type of donorSibling HLA identical 11Haplo-identical 2Sibling mismatched 1Unrelated 16Autologous 19

Conditioning regimensChemotherapy alone 39Chemotherapy and total body irradiation 9Other 1

Days from BMT to admission in PICU<30 2731–99 14>100 8Acute graft-versus-host disease grades III or IV 13

Reasons for admission in PICUPulmonary complications 17 (34.7%)Hemodynamic failure 14 (28.6%)Neurologic disease 10 (20.4%)Renal failure 4 (8.2%)Bleeding 3 (6.1%)Allergic reaction 1 (2%)

Table 2 Outcome and degree of critical care of 49 admissions toPICU after BMT. (PICU pediatric intensive care unit, LI lung inju-ry following criteria of Rubenfeld et al. [3], ARDS acute respirato-ry distress syndrome, BMT hematopoietic stem cell transplantationfrom blood or bone marrow)

Days in PICU (median and range) 7 (1–61)Number of patients supported by mechanical 34ventilationDays in mechanical ventilation (median and range) 5 (1–41)Diagnosis of LI 17Diagnosis of ARDS 7Severe hemorrhagic cystitis 8Need for renal replacement therapy 7Diagnosis of infection 33Neutropenia (neutrophil count <0.5×103/l) 23

Organ failure (number of admissions and percentage)Respiratory 34 (69.4%)Cardiovascular 33 (67.3%)Renal 16 (32.7%)Hepatic 28 (57.1%)Neurologic 12 (24.5%)Gastrointestinal 9 (18.4%)Hematologic 40 (81.6%)

Number of organ failures0 failures 4 (8.2%)1 failure 1 (2.0%)2 failures 6 (12.2%)3 failures 14 (28.6%)4 failures 11 (22.4%)5 failures 6 (12.2%)6 failures 6 (12.2%)7 failures 1 (2.0%)

Discussion

We present the retrospective analysis of the outcome andpredictors of survival of a cohort of child recipients ofBMT who required intensive care support. In our series,the proportion of BMT patients requiring PICU admis-

sion is high (29.1%), and 22.5% required MV. These figures are in accordance with previous studies, in which11–40% of BMT recipients required intensive care unit(ICU) admission [4, 6, 7, 8, 9, 10, 11], and 6–33% need-ed MV [3, 6, 10, 16, 17, 18, 19, 20, 21, 22, 23].

Survival of adult BMT patients requiring ICU supportis poor, ranging between 12% and 46% [4, 7, 10, 11, 17],and between 0% and 19% if they require MV [3, 4, 7,11, 16, 17, 18, 23, 24, 25, 26]. These figures are some-what better for children, whose survival rate is 27–55%[6, 8, 9], and 9–44% if they require MV [8, 19, 20, 21,22, 27]. In our cohort, survival 30 days after PICU dis-charge was 29.6%, supporting the effectiveness of criti-cal care support of BMT patients.

Previous reports have tried to establish prognosticfactors predictive of fatal outcome. One series analysedthe outcome of 865 mechanically ventilated patients, andestablished that younger age, a lower APACHE III score,and a shorter time from transplantation to intubationwere statistically associated with a better prognosis [3].In that study [3] there were no survivors among the pa-tients with both lung injury and more than 4 h of vaso-pressor support or those that had sustained hepatic or re-nal failures. The largest pediatric series analyzed 121mechanically ventilated patients and established as majorrisk factors for death respiratory failure as the reason forendotracheal intubation, the presence of pulmonary in-fection, and the impairment of two or more organ sys-tems on the seventh day after intubation [22].

We did not find significant differences in mortality re-lated to the type of transplant, conditioning regimens,days from BMT to admission to the PICU, underlyingdisease, presence of infection or neutropenia, length ofstay in the PICU, presence of ARDS and presence of renal, neurological, hematological or hepatic failures.Some of these factors are known to influence the clinicalcourse of BMT but they did not affect the outcome ofour patients.

94

Table 3 Correlation of cate-gorical risk factors with surviv-al of 49 admissions to PICU af-ter BMT. (aGVHD acute graftversus host disease, LI lung in-jury following criteria ofRubenfeld et al. [3], ARDSacute distress respiratory syn-drome)

Risk factor Survivors Non-survivors P value OR (95% CI)n=18 n=31

aGVHD III-IV 1 12 0.01 0.1 (0.0–0.78)LI 3 14 0.04 0.25 (0.04–1.15)ARDS 1 6 NS 0.25 (0.01–2.37)Severe hemorrhagic cystitis 0 8 0.01 0.00 (0.00–0.88)Renal replacement therapy 0 7 0.02 0.00 (0.00–1.08)Infection 9 24 NS 0.3 (0.07–1.21)Neutropenia 7 16 NS 1.66 (0.44–6.52)Renal failure 3 13 NS 0.28 (0.04–1.31)Hepatic failure 7 21 NS 0.31 (0.08–1.19)Neurologic failure 3 9 NS 0.5 (0.07–2.44)Hematologic failure 12 28 NS 0.22 (0.03–1.24)Gastrointestinal failure 0 9 0.01 0.00 (0.00–0.74)Cardiovascular failure 8 25 0.009 0.2 (0.04–0.82)Respiratory failure 8 26 0.003 0.16 (0.03–0.69)Male/female 4/14 17/14 0.02 0.23 (0.07–0.83)

Fig. 1 Number of admissions to PICU after BMT related with or-gan systems failure and mortality

Table 4 Correlation of continuous variables with survival of 49admissions to PICU after BMT. Data are median and range.(PICU pediatric intensive care unit)

Risk factor Survivors Non-survivors P value

Age on PICU admission 8 (1–17) 12 (1–18) 0.03(years)Days in PICU 9 (2–61) 6 (1–28) NSNumber organ failures 2 (0–6) 4 (2–7) 0.0001

Some clinical characteristics associated with a lowerchance of survival were identified in our series. The mainrisk factors related to mortality obtained in univariateanalysis were presence of aGVHD grades III or IV, malegender, older age, need for MV, presence of LI, presenceof severe hemorrhagic cystitis, need for renal replacementtherapy, presence of respiratory, cardiovascular or gastro-intestinal failures, and failure of ≥3 organ systems.

Male gender, aGVHD grades III or IV, severe hemor-rhagic cystitis and failure of ≥3 organ systems were in-dependent risk factors related to mortality in the multi-variate analysis. Multiorgan failure was also found inmost series as a predictor of death [7, 19, 20, 22]. Thepresence of aGVHD grades III or IV involves a poor out-come when this appears after BMT. Several pediatric re-ports have also associated this complication with signifi-cantly higher mortality when intensive care is required[8, 9, 20]. Presence of severe hemorrhagic cystitis hadnot been previously reported as a prognostic factor influ-encing fatal outcome in this population. Male gender isrelated with bad outcome in our patients and in other pe-diatric series previously reported [20]. An older age isassociated with a worse outcome in BMT recipients andalso with increasing mortality in a BMT population re-quiring intensive care and MV [3, 16, 24, 25]; we foundsimilar results to these. There were no survivors in thosepatients with gastrointestinal failure. These were patientswith severe aGVHD disease with intestinal involvementand massive gastrointestinal hemorrhage. This bad out-come has also been previously reported [17]. Patients re-quiring dialysis or other methods for renal replacementtherapy have a very low chance for survival in our series,as found in other series [3, 17, 19]. PICU admission dur-ing earlier phases of BMT has been reported as a nega-tive prognostic factor by several authors [3, 17]. Wefound a non-significant trend towards a higher mortalityrate in those patients admitted to the PICU between31 days and 99 days after BMT. Long ICU stay [4, 18],allogenic graft [23], mismatched or unrelated donors [16,17, 25], and the presence of infection [17] have been as-sociated with a worse outcome in some studies, but wedid not find such an association. The need for MV, thepresence of LI, and the need for vasopressor support orhemodynamic failure are risk factors for mortality in ourpatients as in most other series [3, 4, 6, 7, 8, 17, 19, 20,22]. Surprisingly, ARDS was not significantly associatedwith mortality in our patients. This was in contrast withthe study where this risk factor was analysed, whichshowed a negative influence in the outcome [22]. Onlyone of the children who met the criteria for ARDS sur-vived. This lack of statistical significance could be dueto the small number of patients fulfilling criteria forARDS.

A high number of pediatric BMT recipients requirePICU support. Physicians who take care of these childrenin the PICU have usually shared the experience that this

is a very challenging patient population because, as wehave shown, they have a high mortality rate. However, itis also true that certain complications are potentially re-versible with intensive life support, and in our experiencea significant percentage of BMT recipients admitted tothe PICU survive. Our results indicate that the require-ment for MV in BMT recipients should not contraindicateadmission to the PICU, because 23.5% of mechanicallyventilated patients survived. The survival for childrenwith intrinsic pulmonary conditions as the reason for MVwas 17.9%, which was significantly lower than the sur-vival of children intubated and mechanically ventilatedwith non-pulmonary conditions (50%); this chance ofsurvival is enough to justify a therapeutic trial of inten-sive care and assisted mechanical ventilation. These ob-servations need to be considered when the PICU admis-sion of a pediatric BMT recipient is ascertained.

At present, there are not enough data to identify thegroup of BMT pediatric patients with survival after MVnear zero, when the decision of transferring the patient tothe PICU is going to be made. It is reasonable to offerintensive care and MV to all who request it, giving theopportunity for diagnostic studies and reversal of organdysfunction.

The ability to predict survival among the patients ad-mitted to the PICU earlier could avoid futile prolongationof treatment in children with a very low probability ofsurvival. The initiation of intensive care support and MVdoes not demand its continuation when there is no realis-tic expectation of improvement in the patient’s condition.We now have evidence that certain clinical parametersthat can be identified early during the PICU stay are asso-ciated with very low chances of survival. Male gender,older ages, the presence of aGVHD grades III or IV, se-vere hemorrhagic cystitis, lung injury, respiratory, cardio-vascular or gastrointestinal failures, the need for MV, theneed for renal replacement therapy, and the presence ofthree or more organ system failures are prognostic factorsthat predict a poor outcome. If the patient does not re-spond to the intensive support measures and shows thepresence of several of these clinical parameters, the limi-tation of invasive care is probably justified.

We think that these factors and others previously re-ported should be known by the patients or their parentswhen BMT is going to be performed and when admis-sion to the PICU is considered in order to establish a re-alistic prognosis. Some authors have proposed guidelinesfor MV after BMT thereby trying to help physicians andfamilies make decisions regarding withholding or with-drawing life support [3]. As more significant prognosticdata become available, we will be obliged to presentthem to the patients or their families and to use that in-formation to optimize the use of the available resources.

Acknowledgement The statistical analysis was performed by Pilar Martín Pérez, Statistical Techniques, “Conserjería de Sani-dad de la Junta de Castilla y León”, Spain.

95

96

References

1. Armitage JO (1994) Bone marrow trans-plantation. N Engl J Med 330:827–838

2. Roberts I (1994) Bone marrow trans-plantation in children: current resultsand controversies. Bone MarrowTransplant 14:197–199

3. Rubenfeld GD, Crawford SW (1996)Withdrawing life support from me-chanically ventilated recipients of bonemarrow transplants: a case for evidencebased guidelines. Ann Intern Med125:625–633

4. Paz HL, Crilley P, Weinar M, Brodsky I(1993) Outcome of patients requiringmedical ICU admission following bonemarrow transplantation. Chest 104:527–531

5. Cahill RA, Spitzer TR, Mazumder A(1996) Marrow engraftment and clini-cal manifestations of capillary leaksyndrome. Bone Marrow Transplant18:177–184

6. Hayes C, Lush RJ, Cornish JM, FootAM, Henderson J, Jenkins I, Murphy P,Oakhill A, Pamphilon DH, StewardCG, Weir P, Wolf A, Marks DI (1998)The outcome of children requiring ad-mission to an intensive care unit fol-lowing bone marrow transplantation.Br J Haematol 102:666–670

7. Torrecilla C, Cortés JL, Chamorro C,Rubio JJ, Galdos P, Domínguez de Villota E (1988) Prognostic assessmentof acute complications of bone marrowtransplantation requiring intensive ther-apy. Intensive Care Med 14:393–398

8. Warwick AB, Mertens AC, Ou Shu X,Ramsay NKC, Neglia JP (1998) Out-comes following mechanical ventila-tion in children undergoing bone mar-row transplantation. Bone MarrowTransplant 22:787–794

9. Díaz de Heredia C, Moreno A, Olivé T,Iglesias J, Ortega JJ (1999) Role of theintensive care unit in children undergo-ing bone marrow transplantation withlife-threatening complications. BoneMarrow Transplant 24:163–168

10. Jackson SR, Tweeddale MG, BarnettMJ, Spinelli JJ, Sutherland HJ, ReeceDE, Klingemann HG, Nantel SH, FungHC, Toze CL, Philips GL, Shepherd JD(1998) Admission of bone marrowtransplant recipients to the intensivecare unit: outcome, survival and prog-nostic factors. Bone Marrow Trans-plant 21:697–704

11. Afessa B, Tefferi A, Hoagland HC,Letendre L, Peters SG (1992) Outcomeof recipients of bone marrow trans-plants who require intensive-care unitsupport. Mayo Clin Proc 67:117–122

12. Pollack MM, Yeh TS, Ruttiman UE,Holbrook PR, Fields AI (1984) Evalu-ation of pediatric intensive care. Crit Care Med 12:376–383

13. Bernard GR, Artigas A, Brigham KL,Carlet J, Falke K, Hudson L, Lamy M,Legall JR, Morris A, Spragg R and theConsensus Committee (1994) TheAmerican-European consensus confer-ence on ARDS. Definitions, mecha-nisms, relevant outcomes and clinicaltrial coordination. Am J Respir CritCare Med 149:818–824

14. Glucksberg H, Storb R, Fefer A, Buckner CD, Neiman PE, Clift RA,Lerner KG. Thomas ED (1974) Clini-cal manifestations of Graft-versus-hostdisease in human recipients of marrowfrom HLA-matched sibling donors.Transplantation 18:295–304

15. Wilkinson JD, Pollack MM, Glass NL,Kanter RK, Katz RW, Steinhart CM(1987) Mortality associated with multi-ple organ system failure and sepsis inpediatric intensive care unit. J Pediatr111:324–328

16. Crawford SW, Petersen FB (1992)Long-term survival from respiratoryfailure after marrow transplantation formalignancy. Am Rev Respir Dis145:510–514

17. Price KJ, Thall PF, Kish SK, ShannonVR, Andersson BS (1998) Prognosticindicators for blood and bone marrowtransplant patients admitted to an inten-sive care unit. Am J Respir Crit CareMed 158:876–884

18. Denardo SJ, Oye RK, Bellamy P (1989)Efficacy of intensive care for bonemarrow transplant patients with respi-ratory failure. Crit Care Med 17:4–6

19. Rossi R, Shemie SD, Calderwood S(1999) Prognosis of pediatric bonemarrow transplant recipients requiringmechanical ventilation. Crit Care Med27:1181–1186

20. Nichols DG, Walker LK, Wingard JR,Bender KS, Bezman M, Zahurak ML,Piantadosi S, Frey-Simon M, RogersMC (1994) Predictors of acute respira-tory failure after bone marrow trans-plantation in children. Crit Care Med22:1485–1491

21. Todd K, Wiley F, Landaw E, GajewskiJ, Bellamy PE, Harrison RE, Brill JE,Feig SA (1994) Survival outcomeamong 54 intubated pediatric bonemarrow transplant patients. Crit CareMed 22:171–176

22. Keenan HT, Bratton SL, Martin LD,Crawford SW, Weiss NS (2000) Out-come of children who require mechani-cal ventilatory support after bone mar-row transplantation. Crit Care Med28:830–835

23. Martin C, Maraninchi D, Saux P, BlaiseD, Blache JL, Gouin F (1990) Need formechanical ventilation in the treatmentof acute infections complicating bonemarrow trasnplantation. Crit Care Med18:120–121

24. Faber-Langendoen K, Caplan AL,McGlave PB (1993) Survival of adultbone marrow transplant patients re-ceiving mechanical ventilation:a casefor restricted use. Bone Marrow Trans-plant 12:501–507

25. Crawford SW, Schwartz DA, PetersenFB, Clark JG (1988) Mechanical venti-lation after bone marrow transplanta-tion. Risk and clinical outcome. Am Rev Respir Dis 137:682–687

26. Hennessy BJ, White M, Crotty GM(1997) Predicting death in mechanical-ly ventilated recipients of bone marrowtransplants. Ann Intern Med 127:88

27. Bojko T, Notterman DA, GreenwaldBM, De Bruin WJ, Magid MS, GodwinT (1995) Acute hypoxemic respiratoryfailure in children following bone mar-row transplantation:an outcome andpathologic study. Crit Care Med23:755–759