Predictors of mortality in patients undergoing autologous hematopoietic cell transplantation admitted to the intensive care unit

Article metrics


Between January 2001 and July 2006, 1013 patients received autologous hematopoietic cell transplants (AHCT) at Canada's largest transplant center. In this retrospective cohort study of AHCT patients admitted to the intensive care unit (ICU), we describe the outcomes following ICU admission and the variables measured in the first 24 h of ICU admission associated with overall ICU mortality. Results indicate a 3.3% ICU admission rate (n=34) with 13 deaths (1% overall mortality rate, 38% in ICU mortality rate). The worst outcome was in AL amyloid patients of whom 28% were admitted to the ICU, with an ICU mortality rate of 55%. The Sequential Organ Failure Assessment (SOFA) score and Acute Physiology and Chronic Health Evaluation (APACHE II) score in the first 24 h were statistically associated with mortality by univariate analysis. Other variables measured at 24 h and associated with ICU mortality included multiorgan failure, mechanical ventilation, inotropic support >4 h and Gram-negative sepsis. Our data indicate that ICU admission in the autotransplant population is rare and that it is influenced by underlying diagnosis, with AL amyloid patients having the highest risk. Our observations may assist clinical decision-making regarding the continuation of intensive care delivered 24 h after ICU admission.


Allogeneic or autologous hematopoietic SCT (HSCT) is effective treatment for patients with a variety of hematological and other malignancies. Much of the published critical care literature related to HSCT is focused on outcomes after allotransplants because of an increased risk of mortality compared with autotransplants,1 due to complications including GVHD, veno-occlusive disease and infection from long-term immunosuppression.2 Not surprisingly, older studies that included patients receiving either allogeneic or autologous transplants emphasized an overall poor ICU outcome for marrow transplant recipients.3, 4, 5, 6, 7 These studies, however, cannot be extrapolated to autologous transplant recipients as the latter are not at risk for many of the life-threatening complications.

Autologous hematopoietic cell transplants (AHCT) represent approximately 70% of all HSCTs.8 Despite the increased tolerability of an autologous compared with an allogeneic HSCT, ICU admission may represent a significant adverse event, although only a small body of literature exists to help guide clinicians as to which patients benefit from intensive care.9, 10, 11, 12, 13 In this study, we describe the outcome of AHCT recipients transferred to the ICU and identify factors associated with mortality.

Materials and methods

A retrospective medical record review was conducted of all AHCT recipients from Princess Margaret Hospital (PMH) who required intensive care support within 100 days post transplant between January 2001 and July 2006. Indications for AHCT included multiple myeloma, non-Hodgkin's lymphoma, Hodgkin's lymphoma, AL amyloidosis, AML, T-cell lymphoma, testicular malignancies, Waldenstrom's macroglobulinemia and POEMS. Intensive therapy regimens consisted of melphalan (140–180 mg/m2 i.v.) and etoposide (60 mg/kg i.v.) with or without single dose (500 cGy at 50 cGy/min) or fractionated (1200 cGy in 6) TBI for hematological malignancies, except for multiple myeloma and amyloidosis, which were treated with melphalan 140 mg/m2 i.v. All patients undergoing AHCT had full resuscitative directives.

The study was approved by the institutional review boards at the University Health Network and Mount Sinai Hospital. Patients requiring intensive care support were transferred to the Mount Sinai Hospital ICU, the adjoining teaching hospital to PMH. The Mount Sinai Hospital ICU is managed by critical care staff, fellows, and internal medicine house staff and maintained a registered list of autotransplant patients transferred to the unit during the study period.

Demographic and biochemical data collected before patient transfer included: age, gender, disease entity, conditioning regimen, number of CD4 cells infused, transplant date, infection history before transfer, and blood work focusing on hematopoietic (hemoglobin, neutrophil count, platelet count), renal and hepatic indices. Similar data were collected from the Mount Sinai Hospital ICU records, including primary-admitting diagnosis, cumulative length of stay, date of death or discharge, need for mechanical ventilation (MV) or inotropic support within the first 24 h of arrival at the ICU, duration of MV and infections cultured during admission. Vital signs included mean arterial pressure, respiratory rate and temperature (if axillary temperature was recorded, 1 °C was added to the value). The worst hematopoietic, renal and liver parameters during the first 24 h were obtained from patient records, allowing for the calculation of the Sequential Organ Failure Assessment (SOFA) score14 and the Acute Physiology and Chronic Health Evaluation (APACHE II) score.15 Patients who required several admissions to the ICU had data included only from their first admission for statistical analysis. Cause of death was documented by the chief cause of death on each patient's death certificate.

Organ failure during the first 24 h of ICU admission was documented if patients had one or more of the following: respiratory failure, defined as hypoxemia (paO2 <50 mm Hg) or hypercapnea (pCO2 >50 mm Hg), or the need for MV (endotracheal intubation of >24 h duration with positive pressure ventilation); cardiovascular failure, defined as a malignant arrhythmia requiring direct current cardioversion, or need for vasopressor support (dopamine, norepinephrine, epinephrine) for >4 h during the initial 24 h of admission; renal failure, defined as a 50% increase in creatinine level over baseline, an acute rise to >177 μmol/l or the need for acute hemodialysis; liver failure if serum bilirubin >68 μmol/l; and neurologic failure defined as patients with a decreased level of awareness or Glasgow Coma Scale <8. Sepsis was defined as a patient with documented systemic inflammatory response syndrome with a documented infection. Febrile neutropenia was defined as an ANC <0.5 with a temperature >38.5 °C.

Patients discharged from the ICU to a hospital ward at PMH were considered survivors. Prognostic factors related to in-hospital mortality were studied in the complete cohort.

Statistical methods

Descriptive data are presented as percentages, as means±s.d. for normally distributed variables or as medians with their corresponding interquartile ranges (IQR) for non-normally distributed variables. Univariable logistic regression was used to identify associations between continuous variables and mortality in AHCT recipients transferred to the ICU, whereas Fisher's exact test was used for categorical variables. Because of the small sample size, exact logistic regression was performed to determine whether underlying diagnosis, inotropic support, and respiratory failure were associated with mortality. All statistical tests were two-tailed and considered statistically significant at α <0.05. The SAS System for Windows version 9.1 (SAS Institute Inc., Cary, NC, USA) was used for all analyses.


Of the 1013 patients who received autotransplants between January 2001 and July 2006, 34 (3.3%) required ICU admission within 100 days of hematopoietic cell infusion. Median age at admission was 57.0 years (interquartile/IQR 52.0, 62.0; range 28–71), and 53% were female. Indications for AHCT included multiple myeloma (61%), non-Hodgkin's lymphoma (19.6%), Hodgkin's lymphoma (11.1%), AL amyloidosis (3.8%), AML (1.7%), and other malignancies (T-cell angioimmunoblastic lymphoma, germ cell tumors, Waldenstrom's macroglobulinemia). Table 1 indicates the percentage of patients with each diagnosis who required ICU care, along with their respective mortality rates. AL amyloidosis had the highest frequency of admission to the ICU and the worst outcome. Median time to admission post-AHCT was 10.0 days (IQR 5.0, 14.0; range 0–100 days), 8.0 days post-AHCT for non-survivors (IQR 4.0, 10.0; range 1–14 days), and 13.0 days post-AHCT for survivors (IQR 6.0, 16.0; range 0–100 days—see Table 2). Median ICU length of stay was 4.0 days (IQR 2.0, 14.0; range 1–37 days), with a longer length of stay for non-survivors.

Table 1 Percentage admission rate to intensive care unit (ICU) by diagnosis (N=1013)
Table 2 Patient demographics and reason for ICU admission

Reasons for admission to the ICU were often multifactorial, reflecting the potential complications inherent with AHCT. The primary reason for admission as documented in patient charts was systemic inflammatory response syndrome/sepsis in 32% (n=11), followed by respiratory failure in 29% (n=10), and cardiovascular failure in 26% (n=9). Two patients were admitted with upper gastrointestinal hemorrhage. Two patients were admitted for decreased level of consciousness during autograft infusion. The majority of patients with sepsis also developed respiratory failure.

Nine of the patients admitted with sepsis required intubation because of secondary hypoxia or acute respiratory distress syndrome. Reasons for respiratory deterioration requiring urgent intubation during admission included pulmonary edema (n=5), acute respiratory distress syndrome without documented infection (n=3) and peri-engraftment pneumonitis (n=2). Of the nine patients admitted for cardiac-related events, four had pulseless electrical activity (two of which were post-stem cell infusion), three developed unstable arrhythmias and two patients developed hypotension related to autograft infusion requiring vasopressor support. In the first 24 h after admission, 20 patients including 11 non-survivors met the study criteria for respiratory failure, all of whom required >24 h of MV, with a median duration of 3.0 days (IQR 1.0, 11.0; range 1–33 days). The rate of requirement for MV in the whole AHCT population was 2%.

Thirteen of the 34 patients died in the ICU (38% of admissions) with 11 dying from multiorgan system failure, including 9 with concurrent infection. Infections cultured included Escherichia coli bacteremia (3 patients), Candidemia (2 patients), C. difficile (1 patient), Klebsiella (1 patient), Pseudomonal pneumonia (1 patient), and Acinetobacter (1 patient). Other causes of mortality included one multiple myeloma patient dying from a prolonged upper gastrointestinal hemorrhage, and an AL amyloid patient suffering from gross anasarca and intra-alveolar hemorrhage.

In addition to a 3.3% admission rate to the ICU, with a 38% in ICU mortality rate, four patients died on the hospital ward before ICU transfer. These deaths were all secondary to sudden cardiac arrest, which consequently raises the overall in-hospital AHCT mortality rate to 1.7%. Of the 21 survivors who returned to the hospital ward, only one was sent to palliative care (death occurred at day 28 post-discharge), and a second patient died at home on day 33 post-AHCT of unknown causes, resulting in an overall 100 days post-AHCT mortality rate of 1.9%. The mean duration of survival in the survival cohort was 28.8 months (range 28 days to 6 years) post-ICU admission.

Analysis of factors present at 24 h associated with mortality is shown in Table 3. Factors that were associated with mortality included the presence of greater than two organ failure (85% of non-survivors), MV (85% of non-survivors), inotrope dependence (54%) and Gram-negative infection (42%). SOFA and APACHE II scores (calculated among 28 patients), with higher scores reflecting greater degrees of organ dysfunction, were also significantly associated with mortality. Furthermore, patients admitted to the ICU with febrile neutropenia were less likely to die. Of the 15 patients with documented febrile neutropenia during the first 24 h of admission, 13 patients (62% of survivors) survived versus 2 (15% of survivors) that died. Because of the small sample size, underlying diagnosis, MV, and inotropic support were not associated with mortality using exact logistic regression.

Table 3 Variables associated with mortality at 24 h after admission

Additional factors not associated with increased mortality included patient age, number of stem cells infused and time post-AHCT. Biochemical values not associated with mortality included serum lactate, hepatic indices including bilirubin and evidence of renal organ failure. Patients on hemodialysis before AHCT did not have a significantly higher incidence of death; two non-survivors and three survivors were dialysis-dependent pre-transplant.


This study reviewed 34 autotransplant recipients requiring ICU care among a total of 1013 transplants during a 5.5-year period. The admission rate of 3.3% is lower than previous reports that range from 21–44%,3, 4, 5, 6 and even lower if the literature is restricted to the past 10 years, with AHCT admission rates of 8.6–14.6%.11, 12, 13The ICU mortality rate was 38%, representing 1% of all AHCT patients. Patients with AL amyloidosis had the worst outcome, and univariable logistic regression showed that, within the first 24 h, severity of illness scores, >2 organ failures, need for MV or inotropes, and Gram-negative sepsis were associated with increased in-ICU mortality.

Recent literature highlights a trend toward lower admission rates for AHCT patients to the ICU.16 Specifically, Khassawneh et al.,9 in the largest study examining a single homogeneous AHCT cohort, focused on the number of patients requiring MV for supportive care during their transplant period. This study quotes a 6% rate of MV in 1301 patients, compared with 2% in our study.

Several reasons help explain the lower admission rate identified in our study. Patients may be more carefully selected for their suitability for a transplant. One may also expect some patients to decline or not to be offered ICU transfer depending on disease severity; however, all patients maintained a full resuscitative status throughout their admission at PMH, with one eventually being transferred to palliative care following an ICU admission. Critical care goal-directed strategies on the transplant ward have evolved and improved, thereby minimizing the need for ICU transfer. Perhaps the most important explanations for the reduced ICU admission rate are those that minimize the risks of a suppressed bone marrow. These include improved antibiotic and antifungal prophylaxis practices and faster engraftment periods with the use of mobilized peripheral blood progenitor cells rather than those obtained from the bone marrow,17, 18 all of which protect against transplant-related complications with faster reconstitution of the bone marrow.

Our study highlights the importance of the underlying diagnosis in rates of ICU admission and mortality. Thirty-nine patients with AL amyloidosis received AHCT in the studied period, with 11 (28%) requiring ICU admission and 6 (15.4% of the overall AL amyloidosis cohort) eventually dying. This is consistent with earlier studies that quote treatment-related mortality rates of 13–23%.19, 20, 21, 22 This is also not unexpected in patients with AL amyloidosis in whom cardiac and respiratory instability are more common. Over half (54%) of the AL amyloid patients in this study were admitted to the ICU with respiratory or cardiac failure, with most patients eventually dying from multisystem organ failure because of sepsis.

We report an association with the need for MV and ICU mortality (11 of 13 non-survivors required MV). Numerous studies highlight the decreased survivorship of mechanically ventilated patients post-HSCT. However, these primarily represented a large allogeneic population.1, 3, 4, 6, 13 In this study, 45% of patients requiring MV for >24 h survived to ICU discharge. This is comparable with reports for autologous transplant recipient subsets at 30–37%.1, 11 Khassawneh et al.9 provided the largest study of mechanically ventilated patients after AHCT, with an in-hospital survival of 26%. This improved trend to survivorship among ventilated AHCT patients may be explained by improved critical care and early goal-directed therapy,23, 24 improved MV techniques and increased use of non-invasive ventilation strategies.25

It is not surprising that increased degrees of organ failure result in higher mortality rates.3, 6, 9, 13 Khassawneh's study revealed a 32% survival rate for mechanically ventilated patients without evidence of organ failure compared with a 6% in-hospital survival rate if patients had lung injury and vasopressor use, or evidence of hepatorenal failure.9 In this study, the APACHE II and SOFA scores both showed significant differences between survivors and non-survivors. Inotropic dependence within 24 h was also statistically significant with increased mortality (7 of 10 patients died), a correlation that has been cited in earlier studies.6, 12, 13, 21 Inotropic support was required for all patients suffering from Gram-negative infection, a variable that was also independently associated with death. The observation of increased mortality associated with Gram-negative bacterial infection is consistent with other observational HSCT series; however, the study populations consisted predominantly of allogeneic recipients.26, 27

Unlike earlier studies, patients admitted to the ICU with febrile neutropenia had a statistically significant survival benefit, a finding not reported earlier in the literature. This is likely to be related to aggressive antibiotic use, growth factor support and rapid engraftment post-AHCT, all of which may improve survival from infection and its secondary complications. This suggests that the aggressive management of febrile neutropenia in the context of AHCT is warranted.

Given the retrospective nature of this study, several confounding factors should be taken into consideration. For example, patients may have been aggressively resuscitated before ICU transfer, making their 24-h blood parameters less reflective of clinical severity. Furthermore, there are no predefined criteria guiding which patients are transferred to the ICU, and so the population may not be entirely homogeneous. Patients refusing ICU transfer were not an issue, as each patient was of full resuscitative status before transfer.

In conclusion, we observed a low frequency of transfer of the AHCT recipients to the ICU. Mortality in the ICU can be predicted in the first 24 h by specific assessment scores (SOFA and APACHE II); specific supportive care requirements: inotropic dependence and need for ventilation, and clinical findings such as Gram-negative sepsis or >2 organ failure. Febrile neutropenia was associated with decreased mortality. Applying such knowledge will allow clinicians and intensivists to predict the need for critical care support and screen for risk factors associated with mortality in the ICU. This information may have direct consequences on the treatment recommendations offered by health-care providers and on the direction of treatment requested by patients and their families.


  1. 1

    Price KJ, Thall PF, Kish SK, Shannon VR, Andersson BS . Prognostic indicators for blood and marrow transplant patients admitted to an intensive care unit. Am J Respir Crit Care Med 1998; 158: 876–884.

  2. 2

    Copeland E . Hematopioetic stem cell transplantation. N Eng J Med 2006; 354: 1813–1826.

  3. 3

    Torrecilla C, Cortes JL, Chamorro C, Rubio JJ, Galdos P, Dominguez de Villota E . Prognostic assessment of the acute complications of bone marrow transplantation requiring intensive therapy. Intensive Care Med 1988; 14: 393–398.

  4. 4

    Afessa B, Tefferi A, Hoogland HC, Letendre L, Peters SG . Outcomes of recipients of bone marrow transplants who require intensive care support. Mayo Clin Proc 1992; 67: 117–122.

  5. 5

    Crawford SW, Schwartz DA, Petersen FB, Clark JG . Mechanical ventilation after marrow transplantation: risk factors and clinical outcomes. Am Rev Respir Dis 1988; 137: 682–687.

  6. 6

    Rubenfeld CD, Crawford SW . Withdrawing life support from mechanically ventilated recipients of bone marrow transplantation: a case for evidence-based guidelines. Ann Intern Med 1996; 125: 625–633.

  7. 7

    Naeem N, Reed MD, Creger RJ, Youngner SJ, Lazarus HM . Transfer of the hematopoietic stem cell transplant patient to the intensive care unit: does it really matter? Bone Marrow Transplant 2006; 37: 119–133.

  8. 8

    Gratwohl A, Passweg J, Baldomero H, Horisberger B, Urbano-Ispizua A . Economics, health care systems and utilization of hematopoietic stem cell transplants in Europe. Br J Haematol 2002; 117: 451–468.

  9. 9

    Khassawneh BY, White P J, Anaissie EJ, Barlogie B, Hiller FC . Outcome from mechanical ventilation after autologous peripheral blood stem cell transplantation. Chest 2002; 121: 185–188.

  10. 10

    Afessa B, Tefferi A, Dunn WF, Litzow MR . Intensive care unit support and acute physiology and chronic health evaluation III performance in hematopoietic stem cell transplant recipients. Crit Care Med 2003; 31: 1715–1721.

  11. 11

    Soubani AO, Kseibi E, Bander JJ, Klein JL, Khanchandani G, Ahmed HP et al. Outcome and prognostic factors of hematopoietic stem cell transplantation recipients admitted to a medical ICU. Chest 2004; 126: 1604–1611.

  12. 12

    Kew AK, Couban S, Patrick W, Thompson K, White D . Outcome of hematopoietic stem cell transplant recipients admitted to the intensive care unit. Biol Blood Marrow Transplant 2006; 12: 301–395.

  13. 13

    Jackson SR, Tweeddale MG, Barnett MJ, Spinelli JJ, Sutherland HJ, Reece DE et al. Admission of bone marrow transplant recipients admitted to an intensive care unit: outcome, survival and prognostic factors. Bone Marrow Transplant 1998; 21: 697–704.

  14. 14

    Vincent JL, Moreno R, Takala J, Willatts A, De Mondonca A, Bruining H et al. The SOFA score to describe organ dysfunction/failure. Intensive Care Med 1996; 22: 707–710.

  15. 15

    Knaus WA, Draper EA, Wagner DP, Zimmerman JE . Apache II: a severity of disease classification system. Crit Care Med 1985; 13: 818–882.

  16. 16

    Cherif H, Martling CR, Hansen J, Kalin M, Björkholm M . Predictors of short and long-term outcome in patients with hematological disorders admitted to the intensive care unit for a life-threatening complication. Support Care Cancer 2007; 15: 1393–1398.

  17. 17

    Stem Cell Trialists' Collaborative Group. Allogeneic peripheral blood stem-cell compared with bone marrow transplantation in the management of hematologic malignancies: an individual patient date meta-analysis of nice randomized trials. J Clin Oncol 2005; 23: 5074–5087.

  18. 18

    Talmadge JE, Reed E, Ino K, Kessinger A, Kuszynski C, Heimann D et al. Rapid immunologic reconstitution following transplantation with mobilized peripheral blood stem cells as compared to bone marrow. Bone Marrow Transplant 1997; 19: 161–172.

  19. 19

    Gertz MA, Lacy MQ, Dispenzieri A, Gastineau DA, Chen MG, Ansell SM et al. Stem cell transplantation for the management of primary systemic amyloidosis. Am J Med 2002; 112: 549–555.

  20. 20

    Goodman HJ, Gillmore JD, Lachmann HJ, Wechalekar AD, Bradwell AR, Hawkins PN . Outcome of autologous hematopoietic cell transplantation for AL amyloidosis in the UK. Br J Haematol 2006; 134: 417–425.

  21. 21

    Sanchorawala V, Skinner M, Quillen K, Finn KT, Doros G, Seldin DC . Long-term outcome of patients with AL amyloidosis treated with high-dose melphalan and stem-cell transplantation. Blood 2007; 110: 3561–3563.

  22. 22

    Skinner M, Sanchorawala V, Seldin DC, Dember LM, Falk RH, Berk JL et al. High-dose melphalan and autologous stem-cell transplantation in patients with AL amyloidosis: a 8-year study. Ann of Int Med 2004; 140: 85–93.

  23. 23

    Rivers E, Nguyen B, Havstad S, Ressler J, Muzzin A, Knoblich B et al. Early goal directed therapy in the treatment of septic shock. N Eng J Med 2001; 345: 1368–1377.

  24. 24

    The Acute Respiratory Distress Syndrome Network. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Eng J Med 2000; 342: 1301–1308.

  25. 25

    Azoulay R, Alberti C, Bornstain C, Leleu G, Moreau D, Recher C et al. Improved survival in cancer patients requiring mechanical ventilatory support: impact of noninvasive mechanical ventilatory support. Crit Care Med 2001; 29: 519–525.

  26. 26

    Putsiaka DD, Price LL, Ueuzian A, Chan GW, Miller KB, Snydman DR . Blood stream infection after hematopoeitic stem cell transplantation is associated with increased mortaliy. Bone Marrow Transplant 2007; 40: 63–81.

  27. 27

    Collin BA, Leather HL, Wingard JR, Ramphal R . Evolution, incidence, and susceptibility of bacterial bloodstream isolates from 519 bone marrow transplants. Clin Infect Dis 2001; 33: 947–953.

Download references


Dr A Keating holds the Gloria and Seymour Chair in Cell Therapy and Transplantation at University Health Network and the University of Toronto.

Author information

Correspondence to M A Trinkaus.

Rights and permissions

Reprints and Permissions

About this article


  • autologous hematopoietic cell transplantation
  • intensive care unit
  • mortality

Further reading