Epidemiology and prognosis of complications related to allogeneic hematopoietic stem cell transplant (HSCT) recipients requiring admission to intensive care unit (ICU) have not been reassessed precisely in the past few years. We performed a retrospective single-center study on 318 consecutive HSCT patients (2009–2013), analyzing outcome and factors prognostic of ICU admission. Among these patients, 73 were admitted to the ICU. In all, 32 patients (40.3%) died in ICU, 46 at hospital discharge (63%) and 61 (83.6%) 1 year later. Survivors had a significantly lower sequential organ failure assessment (SOFA) score, serum lactate and bilirubin upon ICU admission. Catecholamine support, mechanical ventilation (MV) and/or renal replacement therapy during ICU stay, a delayed organ support and an active graft versus host disease (GvHD) significantly worsen the outcome. By multivariate analysis, the worsening of SOFA score from days 1 to 3, the need for MV and the occurrence of an active GvHD were predictive of mortality. In conclusion, the incidence of HSCT-related complications requiring an admission to an ICU was at 22%, with an ICU mortality rate of 44%, and 84% 1 year later. A degradation of SOFA score at day 3 of ICU, need of MV and occurrence of an active GvHD are main predictive factors of mortality.
Allogeneic hematopoietic stem cell transplantation (HSCT) now represents a widely performed treatment for many hematologic malignancies. Survival after HSCT has improved in the past few decades thanks to advances in human leukocyte antigen matching, conditioning regimens and post-transplant supportive care.1, 2 Extended criteria of inclusion for recipients followed and permitted transplants in older patients with multiple comorbidities.3 However, a high rate of severe complications mainly related to toxicity of conditioning regimen, immunosuppression and GvHD still occur. These complications are often life threatening and require admission to the intensive care unit (ICU).
The incidence of admission of HSC transplant recipients to an ICU is not uniformly appreciated, varying from 10 to 40%.4, 5, 6, 7, 8 Early reports indicate that it carries a very high mortality with low survival rates from 20 to 35%7, 9, 10 and even <10% in patients requiring mechanical ventilation (MV).7 Several studies attempted to identify prognostic factors of critically ill HSCT recipients. Most of them demonstrated that pre-ICU variables are not reliable and should not be used to determine suitability for transfer to the ICU.7, 11 In contrast, the level and the number of organ failures were significantly correlated to mortality.7, 11 In such circumstances, aggressive life support has been considered questionable or even futile especially when MV was indicated. Moreover, long-term survival, although improving, remains poor at <20%.7, 8 Substantial advances in management of both transplantation and critical care of HSCT recipients have been however realized in recent years and may significantly modify the prognosis. Indeed, an ICU survival of 32%, and a 1-year and 5-year survival of 19 and 17% was observed in 164 critically ill allogeneic HSCT recipients followed from 2000 to 2011.12 Furthermore, Lengliné et al.13 observed an improvement in 90-day ICU survival comparing cohorts of patients admitted in two different periods: 1997–2003 and 2004–2011 (31–49%). Nonetheless, there is an obvious lack of recent comprehensive data and this prompted us to carry out a 5-year (2009–2013) analysis of a large cohort of HSCT patients admitted to an ICU in order to study the epidemiology and the clinical features of severe post-transplant complications and to reassess the factors of prognosis and mortality.
Patients and methods
In our hospital, an average of 70 allogeneic HSCTs are performed each year. If any of these patients present a severe and life-threatening condition necessitating a prolonged MV or/and hemodynamic support, they are transferred to our ICU and followed simultaneously with the transplantation team. Patient health-care directives are collected whenever HSCT is decided on. Our policy is that the decision to admit to ICU and to intensively treat these patients should be made by both the intensivists and the hematologists and according to the patient and family wishes.
All consecutive allogeneic HSCT adult recipients admitted to our ICU Unit (ICU of Lapeyronie Montpellier University Hospital) between January 2009 and December 2013 were included in the study. The institutional review board approved the study and waived the need for informed consent.
Clinical charts were retrospectively reviewed and age at ICU admission, sex and underlying hematologic diseases were recorded. HSCT parameters including conditioning regimen, source of stem cells, occurrence of an acute GvHD and its grading and corticosteroid treatment, occurrence of a chronic GvHD and infectious complications were collected.
Upon ICU admission, the reasons for admission and the span from transplantation and hospitalization to admission were noted. The delay between HSCT and admission to ICU was classified in one of the three following periods: <30, 30–90 and >90 days. Clinical and biological variables including mean arterial pressure, diuresis, blood urea nitrogen, serum creatinine, albuminemia, lactatemia and serum bilirubin were monitored. The level of consciousness at ICU admission was evaluated by the Glasgow coma score. The severity of the disease was assessed 24 h after admission using the simplified acute physiology score II (SAPS-II)14 and the sequential organ failure assessment (SOFA) scores.15 The occurrence of aplasia or/and GvHD (active: if treated in the 7 days before ICU admission) was also collected.
During the ICU stay, the need for and the duration of noninvasive and invasive MV and/or renal replacement therapy (RRT) and the use of vasoactive drugs were monitored. Therapies instituted during the first 72 h after ICU admission were considered as initial and those instituted thereafter secondary. The severity of an acute kidney injury, when it occurred, was assessed by the RIFLE (Risk, Injury, Failure, Loss and end-stage renal disease (ESKD)) status.16 During the ICU stay, we also assessed evolution of the disease by SOFA and SAPS-II scores at days 3 and 7, by occurrence of complications, and organ support therapies 72 h after ICU admission. ICU length of stay and outcome, including ICU, 30-day, in-hospital, 6-month and 1-year mortality, were also recorded.
We compared survivors and nonsurvivors for all the above variables in order to identify the predictive factors of ICU and 1-year mortality.
The statistical analyses were performed using the SAS software, version 9 (SAS Institute, Cary, NC, USA). We first performed a descriptive analysis by computing the frequencies and the percents for categorial data; means or medians, s.d., quartiles and extreme values for continuous data. We also checked for the normality of the continuous data distribution using the Shapiro–Wilk tests. The univariate analysis was performed using two-tailed Student’s t-test for continuous variables, Fisher’s test and χ2 test for categorial variables or two-tailed Mann–Whitney-Wilcoxon test when appropriate. To analyze the factors associated with the ICU mortality, univariate and multivariate logistic regression models were performed. We used a hierarchical backward selection process to build the multivariate model. At each step, the variable with the least significant P-value was removed until all remaining main effects in the model were significant at P<0.05. The variable interest was in-ICU mortality, and results were expressed as hazard ratios and 95% confidence interval. The proportional hazards assumption was checked for all logistic regression models constructed. Survival curves were generated using the Kaplan–Meier methodology.
A value of P<0.05 was considered significant.
Population and mortality
During the study period, 318 adult allogeneic HSCTs were performed in the Hematologic Transplant Unit of our hospital. Of these patients, 73 were admitted to the ICU for severe complications, leading to an incidence of life-threatening complications at 21.7%.
A total of 27 females and 46 males, with mean age at 55±7 years, were then studied. Patients’ characteristics, underlying hematologic disease and conventional parameters of allogeneic HSCT are summarized in Table 1. Most of the patients (56; 77%) underwent HSCT from peripheral blood, whereas 17 (24%) had HSCT from cord blood. We observed that 23 patients (31.5%) experienced a high-grade (III or IV) acute GvHD, and corticosteroid treatment was necessary in 46 patients (63%). The median time span from HSCT and hospitalization to ICU admission was 108 (47–132; <30 days: 16; 30–90 days: 26; >90 days: 31 patients) and 18 days, respectively. The ICU admissions were frequently motivated by acute respiratory failure (44 patients, 60.3%).
The clinical and biological data upon admission are displayed in Table 2. Seventeen patients (23.3%) presented an active GvHD and 24 (32.8%) were in aplasia. Seventy-one patients (97.2%) were treated for bacterial infections (suspected or proven), 62 (84.9%) for fungal infections and 47 (64.4%) for viral infections. Microbiological documentation is shown in Table 2, underlying that almost half of the patients have a documented viral infection.
In the first 72 h in the ICU, 25 patients (34.2%) were on catecholamine support, 16 (21.9%) were on noninvasive MV, 29 (39.7%) were on invasive MV (8 were switched from noninvasive to invasive MV) and 18 (24.6%) underwent a RRT (Table 2).
After 72 h in the ICU, more patients were on artificial support: 39 (53.4%) were on catecholamine support, 43 (58.9%) were on invasive MV and 31 (42.4%) underwent a RRT (Table 3). Most of the patients experienced additional complications during their ICU course (Table 3). Infectious complications remain the most frequent, occurring in 83.6% of patients. We reassessed the levels of severity scores at day 3 (60 patients were still present in ICU) and day 7 (38 were still present in ICU). A slight increase in SOFA and SAPS-II scores was observed from days 1 to 7, as displayed in Figures 1 and 2.
The median ICU stay was 9 (3–18) days. Thirty-two patients (43.8%) died in ICU, 5 more (37 patients; 50.7%) at day 30 and 9 more (46 patients; 63%) in the hospital after being discharged from the ICU. This brings the overall in-hospital mortality to 63%. The 6-month and 1-year mortality rates were respectively 74 and 83.6%. Patients who survived ICU admission had a 1-year survival rate of 44.4%. At 1 year after ICU, the quality of life of these patients was fair as reflected by the median Karnofsky score of 80 (37.5–90). Only one-fourth of these patients experienced a score of <50 (normal value at 100). Figure 3 represents the Kaplan–Meier curve for 1-year survival after admission to the ICU.
Prognostic analysis of the 73 allogeneic HSCT recipients admitted to the ICU
The univariate analysis of the data between ICU survivors and nonsurvivors is shown in Tables 1, 2, 3, 4, 5. Neither the underlying hematologic disease nor comorbidities and number of chemotherapy lines nor the time span from the HSCT or hospitalization to the ICU admission were of any prognostic value, nor were the transplant characteristics (Table 1). Among severity scores at ICU admission, only SOFA score was significantly predictive of the mortality. Among biological variables studied upon ICU admission, the following were also significantly predictive of mortality: serum lactate and serum bilirubin. In contrast, the level of renal function impairment upon admission, evaluated by serum creatinine, did not alter the ICU outcome (Table 2).
At ICU admission and during ICU course, the need for MV, catecholamine, RRT and an antiviral treatment significantly worsens the prognosis (Tables 2 and 3). A late introduction, beyond at 72 h, of any organ support was significantly associated with a worse survival rate (P<0.01) as was the occurrence of respiratory, kidney or hepatic failures, active GvHD, digestive hemorrhage and neurological troubles (Table 3). Deceased patients have a significant increase in their severity scores from days 1 to 3 and 7 (Figures 1 and 2).
By multivariate analysis, occurrence of an active GvHD, the need of invasive MV and worsening of SOFA score from days 1 to 3 were independently related to mortality. In contrast, ICU admission SOFA score and the need of an antiviral treatment, RRT and other variables that were associated with mortality in the univariate analysis did not remain significant in the multivariate analysis. The adjusted odds ratios for parameters associated with ICU mortality are summarized in Table 4. ICU length of stay was not different between the two groups.
By comparing 1-year surviving and nonsurviving patients using univariate analysis, we found that SOFA score at day 3 (P=0.01) and to a less extent the necessity of MV during ICU stay (not significant, P=0.06) may predict long-term mortality (Table 5).
The aim of this study was to describe the epidemiology, clinical features and outcome of allogeneic HSCT recipients who were admitted to an ICU during the 5-year period (2009–2013). The incidence of severe complications requiring an admission to ICU was ∼21%. The mortality rate for this population was 43.8% in the ICU, 63% when discharged from the hospital and 83.6% after 1 year. Our results demonstrate that the occurrence of an active GvHD, the need for MV and the worsening of organ dysfunction were the main indicators of prognosis.
Previous studies reported a rate of severe complications requiring critical care varying from 11 to 40% in all HSCT recipients.4, 6, 7, 8, 17 In this study, we reviewed over 300 HSCTs; the incidence of severe complications requiring an ICU admission was 21% comparable to that previously reported.7 The main reason for admission to our ICU was severe respiratory failure as nearly two-thirds of our patients required a ventilatory support, in accordance with previous reports.17, 18 Our population was associated with a high critical illness, as reflected by the level of severity scores, and had ICU and in-hospital survivals of 56 and 37%, respectively. The most recent studies reported ICU survival rates of 46%,11 32%,12 and 49%.13 The improvement in ICU survival that we observed may be related to the multidisciplinary (hematologists and intensivists) approach to the management of these patients and also to the better use of organ support.19 However, long-term survival remains poor as 83.6% of our transplant recipients were deceased after 1 year. Similar results have been reported previously,9, 17 with survival rates that culminate at 19%.7, 11, 12 It is noteworthy that a large proportion (35/61) of our ICU survivors were subjected to therapeutic limitations and were not secondary referred back to the ICU for severe complications.
We reassessed prognosis and survival of HSCT recipients by investigation of this recent time period cohort because identification of the factors predictive of ICU mortality may help to determine suitability for transfer to the ICU and to perform an optimal follow-up and management. First, we found that clinical characteristics before ICU admission such as age, comorbidities, type of HSCT, and conditioning regimen did not influence mortality according to previous reports.7, 17 The use of nonmyeloablative conditioning regimen and PBSCs in some of our transplant recipients that would reduce toxicity and immunosuppression did not induce any significant effect on their ICU survival. However, we did not corroborate the fact that a longer elapsed time since transplantation would be predictive of an increased risk of mortality.7, 20, 21 We observed that ICU outcome of HSCT recipients is more dependent on the underlying critical illness than on the transplant variables or characteristics. The univariate analysis identified eight variables that were related to ICU mortality: SOFA score, serum lactate and bilirubin upon ICU admission; MV and the need for catecholamine support and/or RRT, a delayed organ support (72 h after ICU admission) and an active GvHD. However, when we included these parameters in the multivariate analysis, only the following factors remain predictive: the need for MV, worsening of SOFA score from days 1 to 3 and occurrence of an active GvHD. Among these parameters, the need for MV is a striking determinant, as it dramatically increases the risk of mortality. It corroborates the study of Pène et al.7 who also found in a retrospective analysis that the occurrence of an acute respiratory failure requiring MV in HSCT transplant recipients significantly increases the mortality to 82%. Delayed MV after ICU admission worsens the prognosis even more. Our observations may suggest however that survival of mechanically ventilated allogeneic HSCT recipients is improving with consistent ICU survival of >40% as compared with <20% in the past two decades.7, 10, 12 None of renal cardiovascular and liver failure, when separately investigated, correlated with the probability of mortality in our analysis, whereas an elevated bilirubin level has been considered to be a factor predictive of mortality.7 ICU admission SOFA score15 that includes evaluation of respiratory, renal, cardiovascular and liver functions was significantly lower in surviving patients as compared with the nonsurviving patients.11 Indeed, Neumann et al.17 reported that SOFA score15 was predictive of survival when applied on the day of ICU admission in 64 allogeneic HSCT patients. More important, our multivariate analysis demonstrates that the increase in SOFA score15 from ICU day 1 to day 3 significantly predicts the mortality. To the best of our knowledge, sequential assessment of SOFA score15 during the first ICU days to predict outcome has never been evaluated in HSCT recipients. Only, Lamia et al.22 reported that SOFA score on days 1 and 3 was useful in predicting ICU mortality in 92 patients with onco-hematological malignancy in which 11 were HSCT recipients. Based on our results, the reassessment of SOFA score15 on day 3 of ICU stay may help intensivists in the decision-making process in critically ill HSCT patients. Accordingly, a finding of a treatable condition leading to the rapid inversion of organ dysfunction may significantly improve the prognosis. In ICU patients with hematologic malignancies, bacterial infection has been found to be associated with a more severe initial but a more rapidly reversible organ dysfunction and a subsequent lower mortality compared with other complications.23 The other parameter we found that discriminates ICU surviving and nonsurviving patients was the occurrence of an active GvHD either at ICU admission and/or during ICU course. None of the patients with active GvHD survived in the study of Huaringa et al.20 Pène et al.,7 like others,13 also observed that patients with late complications of HSCT in the setting of active GvHD had a significantly lower outcome. An active GvHD mainly necessitates reinforcement of immunosuppression with high-dose corticosteroid, leading to a profound and sustained immunodeficiency and subsequently to high infectious risk and mortality. The occurrence of a digestive hemorrhage secondary to active GvHD and the institution of an antiviral treatment that also reflects a high level of immunosuppression were significantly associated with ICU mortality in our patients. Patients with immunodeficiency and organ failures related to GvHD, especially respiratory failure requiring MV, bears obviously the worst ICU prognosis and their invasive management in ICU may be questionable. In these circumstances, according to our results, a re-evaluation of organ failure assessed by SOFA score,15 3 days after ICU admission, may help the decision with hematologists to avoid futile aggressive care.
Yet, the prognosis of HSCT recipients admitted to the ICU remains rather bad,7, 9, 17 and the occurrence of these severe complications also affects the long-term outcome. Indeed, we observed like others7, 8, 11 a mortality increase from 44% at ICU discharge to 74% at 6 months and 84% after 1 year. However, the 1-year quality of life was satisfactory as reflected by the score of Karnofsky. Among the factors studied, none has been able to predict the long-term prognosis except the observation of a worse outcome with a higher SOFA score at day 3.
We must acknowledge some limitations to our study. First, it was limited to a single center, although it was conducted in a large cohort of HSCT patients. Second, it is a retrospective study that therefore may be affected by various biases. Third, immunosuppressive treatments have changed between individuals regarding differences in time span from transplantation to ICU admission that may represent additional bias in the interpretation of the results.
In summary, the incidence of severe HSCT complications requiring admission to the ICU is far from negligible as it is >20% of the overall HSCT recipients. They are mostly of pulmonary origin. Despite advances in therapeutic strategies and policies, ICU, in-hospital and 1-year mortality remain high. Yet, this poor ICU outcome should not change physicians’ attitudes toward admitting these patients to the ICU but rather incite them to frequent reappraisal of the benefits of intensive care. A degradation of SOFA score at day 3 of ICU, need of MV and occurrence of an active GvHD are main factors predictive of mortality and may be therefore taken into account in therapeutic decision involving both intensivists and hematologists. In order to improve the likelihood of survival in this population, further research is however still needed regarding critical care management and prediction of mortality.
Gooley TA, Chien JW, Pergam SA, Hingorani S, Sorror ML, Boeckh M et al. Reduced mortality after allogeneic hematopoietic-cell transplantation. N Engl J Med 2010; 363: 2091–2101.
Gratwohl A, Baldomero H, Gratwohl M, Aljurf M, Bouzas LF, Horowitz M et al. Quantitative and qualitative differences in use and trends of hematopoietic stem cell transplantation: a Global Observational Study. Haematologica 2013; 98: 1282–1290.
Passweg JR, Baldomero H, Gratwohl A, Bregni M, Cesaro S, Dreger P et al. The EBMT activity survey: 1990-2010. Bone Marrow Transplant 2012; 47: 906–923.
Afessa B, Tefferi A, Hoagland HC, Letendre L, Peters SG . Outcome of recipients of bone marrow transplants who require intensive-care unit support. Mayo Clin Proc Mayo Clin 1992; 67: 117–122.
Jackson SR, Tweeddale MG, Barnett MJ, Spinelli JJ, Sutherland HJ, Reece DE et al. Admission of bone marrow transplant recipients to the intensive care unit: outcome, survival and prognostic factors. Bone Marrow Transplant 1998; 21: 697–704.
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.
Pène F, Aubron C, Azoulay E, Blot F, Thiéry G, Raynard B et al. Outcome of critically ill allogeneic hematopoietic stem-cell transplantation recipients: a reappraisal of indications for organ failure supports. J Clin Oncol 2006; 24: 643–649.
Soubani AO . Critical care considerations of hematopoietic stem cell transplantation. Crit Care Med 2006; 34: S251–S267.
Scales DC, Thiruchelvam D, Kiss A, Sibbald WJ, Redelmeier DA . Intensive care outcomes in bone marrow transplant recipients: a population-based cohort analysis. Crit Care 2008; 12: R77.
Bach PB, Schrag D, Nierman DM, Horak D, White P, Young JW et al. Identification of poor prognostic features among patients requiring mechanical ventilation after hematopoietic stem cell transplantation. Blood 2001; 98: 3234–3240.
Benz R, Schanz U, Maggiorini M, Sebach JD, Stussi G . Risk factors for ICU admission and ICU survival after allogeneic hematopoietic SCT. Bone Marrow Transplant 2014; 49: 62–65.
Townsend WM, Holroyd A, Pearce R, Mackinnon S, Naik P, Goldstone AH et al. Improved intensive care unit survival for critically ill allogeneic haematopoietic stem cell transplant recipients following reduced intensity conditioning. Br J Haematol 2013; 161: 578–586.
Lengliné E, Chevret S, Moreau A-S, Pène F, Blot F, Bourhis JH et al. Changes in intensive care for allogeneic hematopoietic stem cell transplant recipients. Bone Marrow Transplant 2015; 50: 840–845.
Le Gall JR, Lemeshow S, Saulnier F . A new Simplified Acute Physiology Score (SAPS II) based on a European/North American multicenter study. JAMA 1993; 270: 2957–2963.
Vincent JL, Moreno R, Takala J, Willatts S, De Mendonça A, Bruining H et al. The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine. Intensive Care Med 1996; 22: 707–710.
Bellomo R, Kellum JA, Ronco C . Defining and classifying acute renal failure: from advocacy to consensus and validation of the RIFLE criteria. Intensive Care Med 2007; 33: 409–413.
Neumann F, Lobitz O, Fenk R, Bruns I, Köstering M, Steiner S et al. The sepsis-related Organ Failure Assessment (SOFA) score is predictive for survival of patients admitted to the intensive care unit following allogeneic blood stem cell transplantation. Ann Hematol 2008; 87: 299–304.
Diab KJ, Yu Z, Wood KL, Shmalo JA, Sheski FD, Farber MO et al. Comparison of pulmonary complications after nonmyeloablative and conventional allogeneic hematopoietic cell transplant. Biol Blood Marrow Transplant 2012; 18: 1827–1834.
Hilbert G, Gruson D, Vargas F, Valentino R, Gbikpi-Benissan G, Dupon M et al. Noninvasive ventilation in immunosuppressed patients with pulmonary infiltrates, fever, and acute respiratory failure. N Engl J Med 2001; 344: 481–487.
Huaringa AJ, Leyva FJ, Giralt SA, Blanco J, Signes-Costa J, Velarde RE . Outcome of bone marrow transplantation patients requiring mechanical ventilation. Crit Care Med 2000; 28: 1014–1017.
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.
Lamia B, Hellot M-F, Girault C, Tamion F, Dachraoui F, Lenain P et al. Changes in severity and organ failure scores as prognostic factors in onco-hematological malignancy patients admitted to the ICU. Intensive Care Med 2006; 32: 1560–1568.
Vandijck DM, Depuydt PO, Offner FC, Nollet J, Peleman RA, Steel E et al. Impact of organ dysfunction on mortality in ICU patients with hematologic malignancies. Intensive Care Med 2010; 36: 1744–1750.
The authors declare no conflict of interest.
About this article
Platelet number and graft function predict intensive care survival in allogeneic stem cell transplantation patients
Annals of Hematology (2019)
Long-Term Outcome of Patients With a Hematologic Malignancy and Multiple Organ Failure Admitted at the Intensive Care
Critical Care Medicine (2019)
Net reclassification improvement with serial biomarkers and bed-sided spirometry to early predict the need of organ support during the early post-transplantation in-hospital stay in allogeneic HCT recipients
Bone Marrow Transplantation (2019)
Hematology/Oncology and Stem Cell Therapy (2018)
Improved short- and long-term outcome of allogeneic stem cell recipients admitted to the intensive care unit: a retrospective longitudinal analysis of 942 patients
Intensive Care Medicine (2018)