European Journal of Internal Medicine

Transcription

1 European Journal of Internal Medicine 22 (2011) Contents lists available at ScienceDirect European Journal of Internal Medicine journal homepage: Original article Prognosis of patients with haematological malignancies admitted to the intensive care unit: Sequential Organ Failure Assessment (SOFA) trend is a powerful predictor of mortality Daniël A. Geerse a, Lambert F.R. Span b, Sara-joan Pinto-Sietsma c, Walther N.K.A. van Mook a,c, a Department of Intensive Care Medicine, Maastricht University Medical Centre, Maastricht, The Netherlands; currently Department of Internal Medicine, Catharina Hospital, Eindhoven, The Netherlands b Department of Internal Medicine, division of Haematology, Maastricht University Medical Centre, Maastricht, The Netherlands; currently Department of Haematology, University Medical Centre Groningen and University of Groningen, The Netherlands c Department of Internal Medicine, Division of General Internal Medicine, University Hospital Maastricht, Maastricht, The Netherlands; currently Department of Internal Medicine, Academic Medical Centre, Amsterdam, The Netherlands article info abstract Article history: Received 15 June 2010 Received in revised form 29 October 2010 Accepted 11 November 2010 Available online 8 December 2010 Keywords: Haematological malignancy Intensive care Critical care Mortality Prognosis Organ failure Background: The prognosis of patients with haematological malignancies who are admitted to the ICU is generally poor. In order to optimize care, it is important to be able to determine which patients are most likely to benefit from continuation of treatment after ICU admission. Methods: Data of 86 patients with a haematological malignancy consecutively admitted to the ICU of Maastricht University Medical Centre were examined in a retrospective cohort study in order to identify clinically useful prognostic parameters. Results: ICU mortality was 56% and in-hospital mortality was 65%. Non-survivors had higher APACHE-II and SOFA scores compared with survivors (32±8.0 versus 25±6.5 and 11.5±3.1 versus 8.5±3.0, respectively). The mortality rate was significantly higher in patients with an increasing SOFA score ( 2 points) compared with patients with an unchanged or decreasing SOFA score (72% versus 58% and 21%, respectively). Mortality was also higher in patients requiring invasive mechanical ventilation or inotropic/vasopressor therapy. Conclusion: The mortality rate among patients with haematological malignancies who are admitted to the ICU is high and mainly associated with the severity of illness, as reflected by more severe and worsening organ failure and a need for mechanical ventilation or inotropic/vasopressor therapy. Several factors appear to be associated with a poor outcome, but no absolute predictors of mortality could be identified, although the results suggest that changes in the SOFA score during the stay in the ICU can be helpful in the decision making about the continuation or discontinuation of treatment European Federation of Internal Medicine. Published by Elsevier B.V. All rights reserved. 1. Introduction Over recent decades, intensified treatment protocols have improved the prognosis of patients with a haematological malignancy [1], but this has been accompanied by an increase in the incidence of disease and therapy related complications, which often necessitate supportive care in the intensive care unit (ICU). Survival rates are very low in this category of patients, with ICU mortality rates of 38 to 73% and even higher hospital and long-term mortality rates [2 16]. As a result, decision-making around the admission of patients to the ICU and/or the (dis)continuation of ICU care are difficult to make. Abbreviations: ICU, intensive care unit; APACHE-II, Acute Physiology and Chronic Health Evaluation II; SOFA, Sequential Organ Failure Assessment; AML, acute myeloid leukaemia; SCT, stem cell transplantation; OR, odds ratio. Corresponding author. Departments of Intensive Care and Internal Medicine, Maastricht University Medical Centre, P. Debeyelaan 25, 6229 HX Maastricht, The Netherlands. Tel.: ; fax: address: w.van.mook@mumc.nl (W.N.K.A. van Mook). Preferably, these decisions should be based on the prognoses tailored to individual patients. Several prognostic factors have been identified in prior studies among patients with a haematological malignancy, but most of these studies identified groups of patients with moderate to high risk of mortality. In the presence of conditions associated with 100% mortality, it would seem justified not to initiate or to discontinue therapy. Cornet et al. [11] reported 100% mortality in patients with a SOFA score 15 on admission to the ICU, while Lamia et al. [12] and Vandijck et al. [16] reported that a change in the SOFA score was predictive of mortality and could be used to guide clinical decision making in the ICU. The aim of this study was to identify indicators of a poor prognosis, such as (changes in) SOFA scores, in patients with underlying haematological malignancies that were admitted to the ICU. 2. Methods We retrospectively evaluated data from all patients with a haematological malignancy who were admitted to the ICU of Maastricht University /$ see front matter 2010 European Federation of Internal Medicine. Published by Elsevier B.V. All rights reserved. doi: /j.ejim

2 58 D.A. Geerse et al. / European Journal of Internal Medicine 22 (2011) Medical Centre, the Netherlands, between December 1999 and December Ethical approval was not requested for this retrospective study. Patients with a do-not-resuscitate order indicative of poor functional performance and/or poor prognosis were not admitted to the ICU. Patients re-admitted to the ICU after more than 7 days were regarded as new admissions. We reviewed data from patient charts and electronic patient records relating to: demographics, type and activity of malignancy, type and time of initiation of therapy, indication for ICU admission, treatment related complications (e.g. presence and duration of neutropenia, graft versus host disease), clinical parameters (e.g. arterial blood pressure and heart rate), laboratory values, need for mechanical ventilation (invasive/ non-invasive), vasopressor/inotropic use, resuscitation, documentation of infectious disease (blood cultures, bronchoalveolar lavages, and virology studies) and ICU as well as in-hospital survival. Based on these data we calculated for each patient the Acute Physiology and Chronic Health Evaluation II (APACHE-II) score and the Sequential Organ Failure Assessment (SOFA) scores on days 1, 3, 5 and 7 after admission. Disease status was based on the available chart information and the most recent bone marrow biopsy data. Neutropenia was defined as an absolute neutrophil count of b /L [17]. Persistent neutropenia was defined as neutropenia throughout the ICU stay. Aspergillosis was diagnosed based on criteria defined by Ascioglu et al. [18] Statistical analyses Continuous variables were calculated as means±sd. For categorical variables, we calculated total numbers and percentages. We evaluated differences between survivors and non-survivors using Student's T-test for continuous variables and Pearson's χ²-test or Fisher's exact test for categorical variables. We used logistic regression analysis to examine the relationships between multiple variables and ICU survival. Variables that were present in fewer than 10 patients were excluded from the analysis. The Wald test was used to express the significance of the incorporated variables. InordertoevaluatewhetherachangeintheSOFAscoreduringthe first week after admission was predictive of mortality, we divided the patients into three groups: increase, no change or decrease in SOFA score (Table 2). Twenty-three patients (27%) were excluded from the analysis, because only their SOFA score on day 1 was available. These patients had been discharged to the ward (n=13, 15%) or had died (n=10, 12%) within 2 days. A decrease or increase in SOFA score was defined as a 2 point change between day 1 and the last SOFA score in the first week. We used SPSS (SPSS Inc, Chicago IL) for all analyses and considered pb0.05 as statistically significant. Table 1 Patient characteristics and variables divided by ICU survival. Total ICU non-survivors ICU survivors Subgroup mortality n=86 n =48 n=38 56% Age (mean±sd) 52±13 52±13 53±13 Male sex 50 (58%) 28 (58%) 22 (58%) 56% Hematological malignancy Acute leukaemia 33 (38%) 20 (42%) 13 (34%) 61% Other diseases 53 (62%) 28 (58%) 25 (66%) 53% Disease status Active disease 62 (72%) 35 (73%) 29 (71%) 56% Complete remission 24 (28%) 13 (27%) 11 (29%) 54% Therapy Chemotherapy in last 30 days 54 (63%) 29 (60%) 25 (66%) 54% Stem cell transplantation in last year 31 (36%) 19 (40%) 12 (32%) 61% Autologous 11 (13%) 4 (8%) 7 (18%) 36% Allogeneic, myeloablative 9 (10%) 9 (19%) 0 (0%) 100% Allogeneic, non-myeloablative 11 (13%) 6 (13%) 5 (13%) 55% Therapy related complications Neutropenia on ICU admission 48 (56%) 29 (60%) 19 (50%) 60% Neutropenia N21 days 19 (22%) 11 (23%) 8 (21%) 58% Neutropenia persistent on ICU 23 (27%) 15 (31%) 8 (21%) 65% Graft-versus-host disease 14 (16%) 7 (15%) 7 (18%) 50% Chronic immunosuppressive medication 24 (28%) 14 (29%) 10 (26%) 58% Cytomegalovirus reactivation 6 (7%) 6 (13%) 0 (0%) 100% Probable aspergillosis 5 (6%) 2 (4%) 3 (8%) 40% Indication for ICU admission Respiratory failure 30 (35%) 17 (35%) 13 (34%) 57% Sepsis 30 (35%) 19 (40%) 11 (29%) 63% Heart failure 13 (15%) 6 (13%) 7 (18%) 46% Post resuscitation 4 (5%) 2 (4%) 2 (5%) 50% Neurological 3 (3%) 1 (2%) 2 (5%) 33% Other 6 (7%) 3 (6%) 3 (8%) 50% Organ failure scores (mean±sd) APACHE-II score 28.8± ± ±6.5 SOFA score on day ± ± ±3.1 Highest SOFA score in first week 11.5± ± ±3.1 Any SOFA score 15 in first week 18 (21%) 12 (27%) 6 (16%) 67% Invasive mechanical ventilation (within 24 h) 45 (52%) 34 (71%) 11 (29%) 76% PaO2/FiO2-ratio (mmhg, mean±sd) 212± ± ±119 Vasopressor/ inotropic therapy 45 (52%) 34 (71%) 11 (29%) 76% Resuscitation 14 (16%) 12 (25%) 2 (5%) 86% Left ventricular ejection fraction 40% 20 (23%) 11 (23%) 9 (24%) Positive blood culture 30 (35%) 19 (40%) 11 (29%) ICU duration (no. of days, mean±sd) 9.4±13 7.2± ±16 pb0.05. pb0.01. pb0.001.

3 D.A. Geerse et al. / European Journal of Internal Medicine 22 (2011) Results Table 1 shows the patient data. We evaluated 86 ICU admissions involving 75 different patients, 9 of which had 1, and 1 of which had 2 readmissions. The mean age was 52±13 years. Fifty patients (58%) were male. The most common underlying diseases were acute myeloid leukaemia (AML) (33%) and non-hodgkin's lymphoma (28%), and most patients (63%) had received chemotherapy 30 days prior to ICU admission. The most common indications for ICU admission were sepsis (35%) and respiratory failure (35%). On admission, 48 patients (56%) had neutropenia, of whom 35 (73%) had received chemotherapy and/or stem cell transplantation (SCT) in the last 30 days prior to admission. Neutropenia persisted throughout the ICU stay in 23 patients (27%). In 19 patients (22%) neutropenia had been present for 21 days prior to admission. The APACHE-II and SOFA scores on day 1 were 28.8±7.9 and 10.1±3.4., respectively. Invasive and non-invasive ventilation were initiated within 24 h in 45 patients (52%) and 14 patients (16%), respectively. Inotropic and/or vasopressor support was initiated in 14 patients (52%). ICU mortality was 56% and in-hospital mortality was 65%. ICU mortality in all patients admitted to the ICU during the study period was 21.7%. Table 1 shows the differences between patient variables in the ICU survivors and non-survivors. There were no differences in age and sex, but non-survivors had significantly higher APACHE-II and SOFA scores compared with the survivors (31.6±8.0 versus 24.9±6.5 and 11.5±3.1 versus 8.5±3.0, respectively). There were 6 ICU survivors (33%) among the 18 patients with a maximum SOFA score 15. Compared with patients with an unchanged SOFA score (58% mortality), patients whose SOFA score increased by 2 points during the first week had a significantly (pb0.05) higher mortality rate (72%), whereas patients whose SOFA score decreased by 2 points had a significantly lower mortality rate (21%) (Table 2). In general, there was no association between mortality and type and activity of the underlying disease, although mortality was 100% in some small subgroups, such as patients with acute lymphoblastic leukaemia (n=5). These patients had a significantly (pb0.05) higher severity of illness, with a mean APACHE-II score of 36.2±5.1 (versus 28.8±7.8) and a mean SOFA score of 13.0±4.2 (versus 9.9±3.3) and no decline in the SOFA score during the mean ICU stay of 5 days. The mortality rate was also 100% (p=0.004) in patients who had received a myeloablative allogeneic SCT in the year prior to ICU admission (n=9). These patients did not have higher APACHE-II scores, but their SOFA score on day 1 was significantly higher (13±3.4 versus 10±3.2) compared with the other patients. In 3 of these patients complete remission had been achieved before the SCT. In 7 of these 9 patients the indication for ICU admission was sepsis. Mortality was also 100% in 6 patients with reactivation of cytomegalovirus (p=0.032), all of whom had received an allogeneic SCT in the year prior to ICU admission (4 myeloablative and 2 nonmyeloablative SCTs), while 5 had received immunosuppressive medication for graft versus host disease. These patients had higher day 1 SOFA scores (13±3.3 versus 10±3.3) but their APACHE-II scores were not higher. Mortality was not significantly higher in patients with neutropenia on admission, even when neutropenia persisted throughout the ICU stay or had been present for more than 21 days prior to admission. Mortality was not associated with the indication for ICU admission, underlying graft versus host disease, chronic use of immunosuppressive Table 3 Univariate analysis for variables associated with ICU-mortality. Variable Wald OR 95% CI Age Male sex Acute leukaemia Other disease Disease status Complete remission Active disease Chemotherapy in last 30 days Stem cell transplantation in last year Neutropenia on ICU admission Neutropenia N21 days Neutropenia persistent on ICU Graft-versus-host disease Immunosuppressive medication Indication for ICU admission Other Respiratory failure Sepsis APACHE-II score (OR for each additional point) SOFA score on day 1 (OR for each additional point) Highest SOFA score in first week (OR for each additional point) Any SOFA score 15 in first week Invasive mechanical ventilation within 24 h Inotropic/vasopressor therapy Resuscitation Left ventricular ejection fraction 40% Positive blood culture Trend in SOFA score Decreasing SOFA score 1.00 No change in SOFA score Increasing SOFA score OR, odds ratio. CI, confidence interval. pb0.05. pb0.01. p medication, left ventricular ejection fraction 40% or positive blood culture. It was, however, significantly higher in patients in whom invasive mechanical ventilation was initiated within 24 h of admission (pb0.001), in patients treated with inotropic/vasopressor therapy (pb0.001) and in patients who underwent cardiopulmonary resuscitation (p=0.014). Table 3 shows the results of the univariate logistic regression analysis of different variables in relation to ICU mortality. APACHE-II score and SOFA score on day 1 as well as the highest SOFA score in the first week were all associated with higher mortality, as were invasive mechanical ventilation, inotropic/vasopressor therapy and resuscitation. Compared with a decreasing SOFA score, an unchanging or increasing SOFA score was associated with higher mortality (odds ratio 5.11 and 9.75, respectively; Table 3). In order to examine which combination of factors best described the relation to mortality, we constructed a multivariate model (Table 4), including the parameters that were statistically significant on univariate analysis, using the SOFA trend, as this was the organ failure score that was the strongest predictor of mortality. The model was corrected for age and sex. The analysis revealed a significant relationship between mortality and inotropic/vasopressor therapy and mechanical ventilation as well as with the SOFA trend. Both increasing and unchanging SOFA scores Table 2 ICU mortality in subgroups divided by change in SOFA score. SOFA score change Total (n=63) Survivors (n= 31) Mean SOFA on admission Non-survivors (n=32) Mean SOFA on admission Decrease 19 (30%) 15 (48%) (12%) % No change 26 (41%) 11 (35%) (47%) % Increase 18 (29%) 5 (16%) (41%) % Mortality

4 60 D.A. Geerse et al. / European Journal of Internal Medicine 22 (2011) Table 4 Multivariate analysis of factors that were significantly associated with mortality on univariate analysis, using trend in SOFA score. Variables OR 95% CI Wald p-value Inotropic/vasopressor therapy Invasive mechanical ventilation within 24 h Trend in SOFA score Decreasing SOFA score 1.00 No change in SOFA score Increasing SOFA score Resuscitation Male sex Age OR, odds ratio. CI, confidence interval. -2 Log likelihood were associated with higher mortality, compared with decreasing SOFA scores. 4. Discussion Decisions around the initiation or (dis)continuation of intensive care treatment are challenging, especially where they concern critically ill patients with poor predicted outcome. The aim of this study was to identify factors that can support this type of decision making for critically ill patients with a haematological malignancy. Other studies have investigated outcomes and the prognostic value of numerous factors in this group of patients. Our results confirm the previously reported poor prognosis of these severely ill patients, with a high mean APACHE-II score of 29, and ICU mortality of 56% and inhospital mortality of 65%. The outcome in our study population was significantly worse than that in the general ICU population. Most studies reported no association between ICU outcome and the type of underlying haematological malignancy or treatment characteristics [2 4,6,10 14,16], and ICU outcome appears to be mainly determined by the severity of illness as reflected by organ failure scores like the APACHE-II and SOFA score. Although some prior studies reported an association between AML as underlying disease and higher mortality [8,9], this association was not confirmed in the present study. Our results showed a 100% mortality rate in patients with acute lymphoblastic leukaemia, allogeneic SCT and cytomegalovirus reactivation, but these groups were too small to allow reliable statistical analysis. A common finding in these groups was a more severe organ failure, with a mean SOFA score of N13 in all of these groups. Some earlier studies have reported an association between neutropenia and higher mortality [6,8,10,12,15]. Our results add to the evidence that neutropenia, generally, does not significantly increase ICU mortality, even when it has been present during three weeks before admission or when it persists during the ICU treatment. Benoit et al. [6] and Vandijck et al. [16] reported an association between bacteraemia and a favourable prognosis, but we were unable to demonstrate an association with mortality in our study in which we obtained positive blood cultures in 30% of patients. There appears to be a general agreement in most studies, including the present one, that a need for invasive mechanical ventilation is an important prognostic factor of mortality [4,7,8,13,14,16]. Our results show that it is one of the parameters most strongly associated with ICU mortality. Obviously, patients with respiratory distress requiring mechanical ventilation have higher organ failure scores. Clearly, in these frequently immunocompromised patients, a need for mechanical ventilation is strongly indicative not only of respiratory failure due to pulmonary infection, ARDS or sepsis, but also of an increased risk of serious infectious complications, such as ventilator-associated pneumonia. The need for inotropic or vasopressor therapy is another important prognostic factor which contributes to higher organ failure scores. Our results show a strong association between mortality and both a need for inotropic or vasopressor therapy and a need for invasive mechanical ventilation. In contrast to a study by Cornet et al. [11], in which a SOFA score of 15 was associated with 100% mortality, there were several survivors among patients with a SOFA score 15(33%) in our study. This suggests that, in our study population at least, this SOFA score is not an absolute predictor of mortality. Our inability to identify any absolute predictor of mortality argues against routinely denying intensive care to patients with a haematological malignancy, while it supports the common practice in which patients receive a short trial of treatment in the ICU ( unlimited treatment for a limited period of time ). This implies that decisions about (dis) continuation of ICU treatment are best made at some point during a patient's stay in the ICU. It is therefore important to re-evaluate the prognostic factors during the first days after admission. Lamia et al. [12] reported that changes in three different organ failure scores, including the SOFA score, over the first 3 days were independent predictors of inhospital mortality. A larger study by Vandijck et al. [16] also found a relation between increasing organ failure scores and mortality. In the current study, SOFA scores were obtained during the first 7 days in the ICU. As in the previously mentioned studies, the trend of this score was highly predictive of mortality: patients with a decreasing SOFA score ( 2 points) had a significantly better prognosis than patients with an unchanged or increasing SOFA score. Nevertheless, the fact that mortality did not reach 100% in the group with an increasing SOFA score suggests that intensive care support of patients with progressive organ failure cannot be regarded as futile in all individual patients. At the same time, however, it must be noted that unchanged or increasing SOFA scores still outperform all other parameters as markers of an unfavourable prognosis. 5. Conclusions The mortality rate among patients with a haematological malignancy who are admitted to the ICU is high and mainly associated with the severity of illness, as reflected in high and increasing organ failure scores and a need for mechanical ventilation and inotropic or vasopressor therapy. Although we found several factors that appear to be strongly associated with a poor outcome, we were unable to identify any absolute predictor of mortality during the ICU stay. However, the strong association between mortality and increasing and unchanging SOFA scores during ICU stay suggests that these factors should play an important role in clinical decision making. Regular patient re-evaluation and close communication between haematologist and intensivist are essential for providing optimal care in the ICU which is tailored to individual patients and their family. Contributions Daniël Geerse collected, ordered and analysed the patient data, and drafted the initial version of the manuscript with Walther van Mook. Lambert Span participated in the writing of all sections in the manuscript, with special attention for the haematology issues. Sara- Joan Pinto-Sietsma participated in the writing of all the sections in the manuscript, and provided advice regarding statistical analysis. Walther van Mook is the initiator and coordinator of the article. He wrote the initial version of the manuscript together with Daniel Geerse, with special attention to intensive care issues. Learning points Patients with haematological malignancies have high ICU mortality. Mortality is mainly associated with organ failure scores. Failure of SOFA score to decrease during ICU stay was most strongly associated with mortality.

5 D.A. Geerse et al. / European Journal of Internal Medicine 22 (2011) References [1] Brenner H. Long-term survival rates of cancer patients achieved by the end of the 20th century: a period analysis. Lancet 2002;360: [2] Evison J, Rickenbacher P, Ritz R, Gratwohl A, Haberthur C, Elsasser S, et al. Intensive care unit admission in patients with haematological disease: incidence, outcome and prognostic factors. Swiss Med Wkly 2001;131: [3] Massion PB, Dive AM, Doyen C, Bulpa P, Jamart J, Bosly A, et al. Prognosis of hematologic malignancies does not predict intensive care unit mortality. Crit Care Med 2002;30: [4] Kroschinsky F, Weise M, Illmer T, Haenel M, Bornhaeuser M, Hoeffken G, et al. Outcome and prognostic features of intensive care unit treatment in patients with hematological malignancies. Intensive Care Med 2002;28: [5] Silfvast T, Pettila V, Ihalainen A, Elonen E. Multiple organ failure and outcome of critically ill patients with haematological malignancy. Acta Anaesthesiol Scand 2003;47: [6] Benoit DD, Vandewoude KH, Decruyenaere JM, Hoste EA, Colardyn FA. Outcome and early prognostic indicators in patients with a hematologic malignancy admitted to the intensive care unit for a life-threatening complication. Crit Care Med 2003;31: [7] 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: [8] Depuydt PO, Benoit DD, Vandewoude KH, Decruyenaere JM, Colardyn FA. Outcome in noninvasively and invasively ventilated hematologic patients with acute respiratory failure. Chest 2004;126: [9] Bahammam AS, Basha SJ, Masood MI, Shaik SA. Outcome of patients with hematological malignancies admitted to the intensive care unit with life-threatening complications. Saudi Med J 2005;26: [10] Owczuk R, Wujtewicz MA, Sawicka W, Wadrzyk A, Wujtewicz M. Patients with haematological malignancies requiring invasive mechanical ventilation: differences between survivors and non-survivors in intensive care unit. Support Care Cancer 2005;13: [11] Cornet AD, Issa AI, van de Loosdrecht AA, Ossenkoppele GJ, Strack van Schijndel RJ, Groeneveld AB. Sequential organ failure predicts mortality of patients with a haematological malignancy needing intensive care. Eur J Haematol 2005;74: [12] Lamia B, Hellot MF, 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: [13] Ferra C, Marcos P, Misis M, Morgades M, Bordeje ML, Oriol A, et al. Outcome and prognostic factors in patients with hematologic malignancies admitted to the intensive care unit: a single-center experience. Int J Hematol 2007;85: [14] Cherif H, Martling CR, Hansen J, Kalin M, Bjorkholm 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: [15] Vandijck DM, Benoit DD, Depuydt PO, Offner FC, Blot SI, Van Tilborgh AK, et al. Impact of recent intravenous chemotherapy on outcome in severe sepsis and septic shock patients with hematological malignancies. Intensive Care Med 2008;34: [16] 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: [17] Hoffman R, Benz Jr EJ, Shattil SJ, Furie B, Silberstein LE, McGlave P, et al. Hematology: basic principles and practice. Philadelphia: Churchill Livingstone; [18] Ascioglu S, Rex JH, de Pauw B, Bennett JE, Bille J, Crokaert F, et al. Defining opportunistic invasive fungal infections in immunocompromised patients with cancer and hematopoietic stem cell transplants: an international consensus. Clin Infect Dis 2002;34:7 14.