Acute respiratory failure and infectious pneumonia are the major causes of death during induction chemotherapy of acute leukemia. However, the causes, incidence and prognostic value of all respiratory events (REs) occurring in this context have never been assessed prospectively. We recruited 65 consecutive patients with newly diagnosed acute leukemia into a 1-year prospective study (December 2000–November 2001) to evaluate the incidence and prognostic value of these events. REs were frequent: 38 were recorded in 30 patients. There was a significant relationship between REs and pre-existing respiratory disease and/or smoking. REs were caused by infection in 34% of cases, by an established cause other than infection in 42% and had an undetermined cause in 24%. Poor early outcome (death within 45 days of starting induction chemotherapy) in patients experiencing an RE was independently associated with a >25/min respiratory rate (P=0.003) and the nonachievement of complete remission (CR) (P<0.0001). Predictors of overall survival in the entire patient population were the absence of CR (P<0.0001), REs (P=0.02) and a ⩾2 performance status (P=0.03). In conclusion, REs are frequent during induction chemotherapy of acute leukemia and represent an independent prognostic factor of poor outcome, regardless of their cause.
The prognosis of acute leukemia remains poor despite improvements in supportive therapy. Treatment fails in over half of patients. Resistance to treatment occurs in 25–35% of cases and the death rate during aplasia is 15–25%. Moreover, the complications of intensive chemotherapy and, in particular, infections during chemotherapy-induced neutropenia compromise the benefits of curative treatment.1,2 Patients with acute leukemia often experience severe respiratory complications such as acute respiratory failure and infectious pneumonia.3,4 High mortality rates (80–90%) have been reported in patients requiring mechanical ventilation.5
Published studies dealing with pulmonary complications in acute leukemia are retrospective studies and have focused only on pulmonary infiltrates of infectious origin.4,5,6 Since any respiratory event (RE), even when initially minor, may precede severe respiratory complications, we hypothesized that the occurrence of any RE may be of prognostic value. Thus, the aim of this prospective study was to determine the incidence, cause and prognostic value of all REs occurring during induction chemotherapy in patients with acute leukemia.
Patients and methods
This was a prospective study of all consecutive adult patients with acute leukemia diagnosed between December 1, 2000 and November 31, 2001 in the Hematology Department of a university teaching hospital. The diagnosis of acute leukemia was based on the French–American–British (FAB) criteria. Immunophenotyping was performed using flow cytometry.
Induction chemotherapy and first-line antimicrobial treatments
Treatment was as follows: anthracyclines and cytosine arabinoside with or without etoposide for acute myeloblastic leukemia (AML), anthracyclines with or without cytosine arabinoside and all-trans-retinoid acid for acute promyelocytic leukemia, and prednisone, vincristine, cyclophosphamide and anthracyclines for acute lymphoblastic leukemia (ALL). Neutropenic patients with fever received broad-spectrum antibiotic regimens containing a β-lactam and an aminoglycoside. If fever did not resolve within 3–5 days, antifungal treatment (amphotericin B) was initiated. Granulocyte colony-stimulating factor was administrated to patients with ALL from day 15 after the onset of induction chemotherapy until neutrophil recovery. Prophylactic cotrimoxazole was administrated from the onset of chemotherapy until the end of treatment. All neutropenic patients received decontaminated feeding and were isolated in rooms ventilated with HEPA filtration, in which visitors wear nasal masks and gloves.
Definition and diagnosis of REs
An RE was defined as any episode of new respiratory symptom or sign such as dyspnea, cough, sputum, chest pain, rales, hemoptysis, pulmonary infiltrate on chest X-ray or oxygenation impairment. Chest X-ray was performed in all patients with respiratory symptoms or fever. Each RE was confirmed and monitored by a hematologist and a lung specialist. All patients underwent at least three sets of blood culture tests, cytobacteriological analysis of urine, a serum test for Aspergillus antigen (Platellia® Bio-Rad; Marne la Coquette, France). If these investigations failed to identify the cause of the RE or if the chest X-ray was abnormal, a computed tomography (CT) scan of the lungs was performed. Then a spiral CT scan was performed when no pulmonary infiltrate was present or when pulmonary embolism was clinically suspected. Finally, a guided bronchoalveolar lavage (BAL) was performed in pneumonia cases (as defined by the presence of any infiltrate) after obtaining the consent of the patient and of the hematologist in charge. The aspect of the BAL fluid (BALF) was recorded. BALF smears were stained for a differential cell count and to identify usual and opportunistic pathogens (mycobacteria, viruses, fungi, parasites). BALF aliquots were cultured to identify pathogens. When Legionella pneumophila was suspected, BAL was cultured on specific ground and urine was tested for the antigen. Microbiological sputum examination was performed when sputum was available.
A multidisciplinary team (a hematologist, a pulmonologist, a specialist of infectious diseases and a radiologist) reviewed all patients each week, recording RE occurrence and establishing RE cause. REs were classified into three categories according to the criteria given in Table 1: (i) infectious origin; (ii) noninfectious origin; and (iii) undetermined origin.
Prospective data collection
At the time of acute leukemia diagnosis, recorded items were any pre-existing respiratory disease (asthma, chronic obstructive pulmonary disease, tuberculosis), cigarette smoking expressed in pack-years, pre-existing cardiac disease and left ventricular function (as assessed by isotopic ventriculography), any steroid medication, and all established hematological prognostic factors, that is age >/⩽60 years, WHO performance status (PS) ⩾/<2, morphologic subtype, cytogenetic abnormalities,10,11 secondary/de novo acute leukemia, WBC >/⩽20.109/l, P-glycoprotein activity and CD34 expression.
Clinical, biological and radiological data relating to the RE were evaluated, as well as the response to antibacterial and antifungal treatment. Acute respiratory failure was defined as the need for mechanical ventilation (invasive or noninvasive) or a PaO2/FIO2 <200 mmHg.
The prognostic value of REs was assessed on the basis of overall survival and survival rate at day 45 of starting induction chemotherapy. This time point was chosen since it allows to know whether patients have achieved complete hematological remission (bone marrow total blasts <5%, normal blood cell count), and thus if death is related to the RE or to disease progression.
Risk factors for RE and death within 45 days (for patients with RE) of initiating induction therapy were studied by univariate analysis using Student's t-test for continuous variables, the χ2-test for categorical variables and Fisher's exact test for small expected frequencies. Significant differences (P<0.05) observed on univariate analysis were evaluated by logistic regression (multivariate analysis). Prognostic factors were assessed using the log-rank test. Significant prognostic factors (P<0.05) found in univariate analysis were evaluated by multivariate analysis using Cox's proportional-hazards model. We used the Statview software (version 5.0) for statistical analysis (SAS Institute Inc., CA, USA).
Patient characteristics and risk factors for REs
During the 1-year study period, acute leukemia was diagnosed in 67 patients. Two patients were not included because of progressive breast cancer in one case and of HIV infection in the other. Sex ratio was 36/29 (men/women). Median age was 55 years (range, 17–85 years). The number of patients according to leukemia FAB classification was: 39 AML (60%), eight ALL (12%), 11 patients (17%) with acute leukemia secondary to myelodysplastic syndrome and seven patients (11%) with acute leukemia secondary to myeloproliferative syndrome. Median follow-up was 574 days (range, 3–751 days). In all, 30 patients (46%) experienced at least one RE (six patients with two REs, one patient with three REs) during induction chemotherapy. Patient characteristics according to the presence or absence of RE are given in Table 2. Pre-existing respiratory disease and/or smoking was the only factor significantly related to RE occurrence (16/30 vs 6/35, P=0.002; odds ratio: 5.5; 95% confidence interval (CI): 1.8–17.2). Only one of the 12 smokers had chronic obstructive pulmonary disease. Other respiratory disorders included asthma (n=2), tuberculosis (n=1) and chronic restrictive respiratory failure (n=1).
Etiology of REs
The cause of the 38 REs was infectious in 34% of cases, an established cause other than infection in 42% and an undetermined cause in 24%, as detailed in Table 3. Among the nine patients with undetermined etiology, eight (21%) were cured by antibiotics and could be considered as probable bacterial pneumonia, increasing the proportion of infectious causes of REs to 55%. Infection was due to Aspergillus in seven cases (ie an incidence of 10 or 6% if considering only the five proven and probable cases).
Of the 38 REs, 14 were recorded at diagnosis of acute leukemia and 24 occurred after starting induction chemotherapy. All but two REs of undetermined etiology (ie 7/9) were present at diagnosis. REs of noninfectious origin occurred within 10 days in all cases but one (15/16). REs of fungal origin occurred after a median of 17 days (range, 4–30 days). Five of the seven cases of pulmonary aspergillosis occurred after more than 10 days (Figure 1). The incidence of pre-existing respiratory disease or smoking was similar in the infectious and noninfectious groups. There was no relationship between these two groups and left ventricular systolic function impairment, even for REs due to cardiac disease.
Factors predicting poor early outcome of patients with REs
Death within 45 days of starting induction chemotherapy occurred in 18 patients (27.7%). Of the 30 patients with REs, 16 were dead at day 45 (54%). Of these 30 patients, 10 achieved complete remission (CR), 10 did not and 10 died before hematopoietic recovery and evaluation of response to chemotherapy. All of the 10 patients with REs who achieved CR were alive at day 45, while six of the 10 patients with REs who did not achieve CR died before day 45, all these deaths being related to leukemia progression.
Death was clearly related to the RE in the 10 patients who died before hematopoietic recovery and evaluation of response to chemotherapy: two patients with pulmonary aspergillosis, one Candida pneumoniae, three patients with pulmonary bacterial infection and four patients with an RE of noninfectious origin. In the group of patients without RE (n=35), 12 did not achieve a CR including two patients who died of leukemia progression and 23 achieved CR and were alive at day 45 (Figure 2).
In patients with an RE, factors that were significantly related to death by day 45 in univariate analysis were: more than a single pulmonary infiltrate (P=0.01), high respiratory rate (>25/min) (P<0.0001), age >60 years (P=0.003), PS ⩾2 (P=0.0005) and absence of CR (P<0.0001). In multivariate analysis, high respiratory rate (>25/min) (P=0.0025) and absence of CR (P<0.0001) were the only independent determinants of a poor early outcome. Acute respiratory failure occurred in 66% of patients with an RE (20/30); 12 patients were admitted to the intensive care unit and seven died. Need for intensive care and cause of the RE were not related to a poor outcome.
Prognostic factors for overall survival
Univariate analysis indicated that the following factors were correlated with overall survival: age >60 years (P=0.008), secondary acute leukemia (P=0.002), PS ⩾2 (P<0.0001), absence of CR (P<0.0001), cytogenetic abnormalities (P=0.0001) and RE (P=0.007). In multivariate analysis, RE (P=0.02; RR: 3.66; 95% CI: 1.29–10), absence of CR (P<0.0001; RR: 15; 95% CI: 4–56) and PS ⩾2 (P=0.03; RR: 2.5; 95% CI: 1–3) were independent determinants of survival (Figure 3a–c).
This prospective study found a high rate of REs in patients with newly diagnosed acute leukemia. REs occurred in 46% of the 65 patients, especially in smokers or patients with previous respiratory disease. REs were due to infection in 34% of cases and to another cause in 42%. Irrespective of their cause, REs were of a major prognostic significance.
The 24% rate of REs with undetermined etiologies agrees with a recent study of immunocompromised patients.12 However, infectious etiologies and especially proven infections were less frequent in our patients than in that study (34 vs 58%).12 This discrepancy may be related to the markedly heterogeneous immunocompromising conditions in that study (in which only 27% of patients were neutropenic) and to the lower proportion of our patients who underwent FOB (37 vs 67% in Rano's study).12 Our low rate of FOB can be explained by the low bacteriological diagnostic yield of FOB in neutropenic AL patients undergoing induction chemotherapy and receiving early broad-spectrum antibiotics,13 and the potentially life-threatening complications of FOB in patients with severe thrombocytopenia, neutropenia and hypoxemia.14 Comparable rates of FOB were reported by Wilhelm et al6 (25%) and Ewing et al5 (43%) in neutropenic patients with acute leukemia. Among infectious etiologies, we found a 6% incidence of proven and probable pulmonary aspergillosis, which agrees with the recently published rates in AML and ALL (8 and 6.3%, respectively).15
Half of the REs of noninfectious origin in our study were related to pulmonary edema. All occurred during the first week after the onset of induction treatment, suggesting that chemotherapy and hydration were involved. They were not predicted by pretreatment isotopic measurement of left ventricular ejection fraction, but this investigation does not allow to rule-out diastolic myocardial dysfunction. In this context, the combination of high volumes of hydration, drugs such as anthracyclines and cytosine arabinoside and anemia could impair cardiopulmonary function. Alternatively, we cannot rule out transient noncardiogenic pulmonary edema in some patients, because no measurement of pulmonary capillary wedge pressure or extravascular lung water was performed.
The usual hematological factors predicting outcome for AL were not different in patients with or without REs. Conversely, REs were significantly more frequent in smokers (>20 pack-year) or patients with pre-existing lung disease. There may be several explanations for the increased rate of REs in smokers: (i) cigarette smoke might impair the clearance and detoxification of bacteria, resulting in a higher prevalence of bacterial colonization of proximal or distal airways;16,17 (ii) smoking is associated with immunologic disturbances;18,19 (iii) in healthy long-term smokers, myocardial blood flow is impaired and the degree of impairment correlates with the number of years the patient has smoked;20,21 and (iv) in animal models, cigarette smokes impairs cardiopulmonary function and increases capillary permeability. This could explain the hydration-related pulmonary edema occurring at the onset of chemotherapy.22,23 Unlike Rossini et al,4 we did not find that total blood count, blast count and age were associated with more frequent pneumonia.
The presence of an RE during induction chemotherapy was a significant prognostic factor, independent of usual hematological factors. If an RE occurred, hematological factors (no CR) and respiratory rate >25/min were predictors of death by day 45. Two-thirds of the patients with an RE developed acute respiratory failure, suggesting that each RE should be closely monitored by a hematologist and a lung specialist. Admission to the intensive care unit for acute respiratory failure was not significantly associated with a poor outcome, which could be explained by (i) the selection criteria used to decide whether patients should be admitted to the unit (age, disease response to treatment, cytogenetic abnormalities and secondary acute leukemia) and/or (ii) the use of noninvasive mechanical ventilation as first ventilatory support, which improves the prognosis of ICU-admitted patients.24
Outcome did not depend on the cause of the RE (infectious or noninfectious) nor on the type of infection, despite previous demonstration of a poorer prognosis in patients with neutropenia-related fungal infections.6 Our low mortality rate may be due to the weekly screening for aspergillosis antigenemia and to earlier introduction of antifungal treatment (within 3–5 days) than in most institutions (5–7 days) when fever did not resolve with broad-spectrum antibiotics. In addition, two of our seven patients with fungal infections had localized pulmonary aspergillosis and were cured by surgery.
Improvement of survival in acute leukemia needs achievement of CR. In fact, remission also appears to be crucial for a favorable outcome of the RE since it implies resolution of neutropenia. Improving survival also requires a better initial management of patients with an RE. This could be achieved by identifying patients at a higher risk of RE and using antimicrobial prophylaxis in smokers and patients with pre-existing pulmonary disease,25,26 earlier systemic antifungal therapy in patients with fungal infection, monitoring of hydration to avoid pulmonary edema and earlier diagnosis of pneumonia by CT scanning in neutropenic patients with unexplained fever for >48 h.27
In conclusion, clinically defined REs are frequent and occur early in the course of acute leukemia. Half are of infectious origin and all are of a major prognostic significance. This stresses the importance of early diagnosis and appropriate management. These results should be confirmed in a multicenter study in which at-risk patients are closely monitored for RE.
Lowenberg B, Dowing JR, Burnett AK . Acute myeloid leukemia. New Engl J Med 1999; 341: 1051–1061.
Baudard M, Beauchamp-Nicoud A, Delmer A, Rio B, Blanc CM, Zittoun R et al. Has the prognosis of adult patients with acute myeloid leukemia improved over years. A single institution experience of 784 consecutive patients over a 16-years period. Leukemia 1999; 13: 1481–1490.
Krowka MJ, Rosenow EC, Hoagland HC . Pulmonary complications of bone marrow transplantation. Chest 1985; 87: 237–246.
Rossini F, Verga M, Pioltelli P, Giltri G, Sancassani V, Pogliani EM et al. Incidence and outcome of pneumonia in patients with acute leukemia receiving first induction therapy with anthracyclines-containing regimens. Haematologica 2000; 85: 1255–1260.
Ewig S, Glasmacher A, Ulrich B, Wilhelm K, Schafer H, Nachtsheim KH . Pulmonary infiltrates in neutropenic patients with acute leukemia during chemotherapy. Chest 1998; 114: 444–451.
Wilhelm M, Kantarjian HM, O’Brien S, Pierce S, Keating MJ, Freireich EJ et al. Pneumonia during remission induction chemotherapy in patients with AML or MDS. Leukemia 1996; 10: 1870–1873.
Cordonnier C, Escudier E, Verra F, Brochard L, Bernaudin JF, Fleury-Feith J . Bonchoalveolar lavage during neutropenic episodes: diagnostic yield and cellular pattern. Eur Respir J 1994; 7: 114–120.
Ewig S, Torres A, Riquelme R, El-Elbiary M, Rovira M, Carreras E et al. Pulmonary complications in patients with haematological malignancies treated at a respiratory ICU. Eur Respir J 1998; 12: 116–122.
Ascioglu S, Rex JH, Pauw B, Bennett JE, Bille J, Crokaert F et al. Defining opportunistic invasive fungal infections in immunocompromised patients with cancer and hematopietic stem cell transplants: an international consensus. Clin Infect Dis 2002; 34: 7–14.
Grimwalde D, Walker H, Oliver F, Wheatley K, Harrison H, Rees J et al. The importance of diagnostic cytogenetics on outcome in AML: analysis of 1612 patients entered into the MRC AML 10 trial. Blood 1998; 92: 2322–2333.
Groupe Français de Cytogénétique Hématologique. Cytogenetic abnormalities in adult acute lymphoblastic leukemia: correlations with hematologic findings and outcome. Blood 1996; 87: 3135–3142.
Rano A, Agusti C, Jiminez P, Angrill J, Benito N, Danès C et al. Pulmonary infiltrates in non-HIV immunocompromised patients: a diagnostic approach using non-invasive and bronchoscopic procedures. Thorax 2001; 122: 253–261.
Saito H, Anaissie EJ, Morice RC, Dekmezian R, Bodey GP . Bronchoalveolar lavage in the diagnosis of pulmonary infiltrates in patients with acute leukemia. Chest 1994; 94: 745–749.
Mayaud C, Cadranel J . A persistent challenge: the diagnosis of respiratory disease in the non-AIDS immunocompromised host. Thorax 2000; 55: 511–517.
Cornet M, Fleury L, Maslo C, Bernard G, Brüker G, the Invasive Aspergillosis Surveillance Network of the Assistance Publique-Hôpitaux de Paris. Epidemiology of invasive aspergillosis in France: a six-year multicentric survey in the greater Paris area. J Hosp Infect 2002; 51: 288–296.
Irwin RS, Erickson AD, Pratter MR, Corrao WM, Garrity FL, Kaemmerlen JT . Prediction of tracheobronchial colonization in current cigarette smokers with chronic obstructive bronchitis. J Infect Dis 1982; 145: 234–241.
Zalacain R, Sobradillo V, Amilibia J, Barron J, Achotegui V, Llorente JL . Predisposing factors to bacterial colonization in chronic obstructive pulmonary disease. Eur Respir J 1999; 13: 343–348.
Corre F, Lellouch J, Schwartz D . Smoking and leucocyte-counts. Results of an epidemiological survey. Lancet 1971; 2: 632–634.
Tollerud DJ, Clark JW, Brown LM, Mann DL, Neuland CY, Pankiw-Trost LK et al. The effects of cigarette smoking on T cells subsets. A population-based survey of healthy Caucasians. Am Rev Respir Dis 1989; 139: 194–198.
Campisi R, Czernin J, Schöder H, Sayre JW, Marengo FD, Phelps ME et al. Effects of long-term on myocardial blood flow coronary vasomotion, and vasodilatator capacity. Circulation 1998; 98: 119–125.
Madsen MM, Bottcher M, Toftegaard Nielson T, Czernin J . Altered regulation of the microcirculation in younger smokers. Cardiology 2000; 94: 91–98.
Gandini C, Castoldi AF, Candura SM, Locatelli C, Butera R, Priori S et al. Carbon monoxide cardiotoxicity. J Toxicol Clin Toxicol 2001; 39: 35–44.
Wang CZ, Li A, Yang ZC . The pathophysiology of carbon monoxide poisoning and acute respiratory failure in a sheep model with smoke inhalation injury. Chest 1990; 97: 736–742.
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.
Menichetti F, Del Favero A, Martino P, Bucaneve G, Micozzi A, Girmenia C et al. Itraconazol oral solution as prophylaxis for fungal infections in neutropenic patients with hematologic malignancies: a randomized, placebo-controlled, double-blind multicentric trial. Clin Infect Dis 1999; 28: 250–255.
Morgenstern GR, Prentice AG, Prentice HG, Ropner JE, Schey SA, Warnock DW . A randomized controlled trial of itraconazole versus fluconazole for the prevention of fungal infections in patients with haematological malignancies. Br J Haematol 1999; 105: 901–911.
Heussel CP, Kauczor HU, Heussel GE, Fisher B, Begrich M, Mildenberger P et al. Pneumonia in febrile neutropenic patients and in bone marrow and blood stem-cell transplant recipients: use of high-resolution computed tomography. J Clin Oncol 1999; 17: 796–805.
About this article
Cite this article
Chaoui, D., Legrand, O., Roche, N. et al. Incidence and prognostic value of respiratory events in acute leukemia. Leukemia 18, 670–675 (2004). https://doi.org/10.1038/sj.leu.2403270
- acute leukemia
- pulmonary complications
The utility and safety of flexible bronchoscopy in critically ill acute leukemia patients: a retrospective cohort study
Canadian Journal of Anesthesia/Journal canadien d'anesthésie (2018)
Prise en charge du patient neutropénique en réanimation (nouveau-nés exclus). Recommandations d’un panel d’experts de la Société de réanimation de langue française (SRLF) avec le Groupe francophone de réanimation et urgences pédiatriques (GFRUP), la Société française d’anesthésie et de réanimation (Sfar), la Société française d’hématologie (SFH), la Société française d’hygiène hospitalière (SF2H) et la Société de pathologies infectieuses de langue française (SPILF)
Médecine Intensive Réanimation (2017)
memo - Magazine of European Medical Oncology (2016)
Increased mortality in hematological malignancy patients with acute respiratory failure from undetermined etiology: a Groupe de Recherche en Réanimation Respiratoire en Onco-Hématologie (Grrr-OH) study
Annals of Intensive Care (2016)
Management of neutropenic patients in the intensive care unit (NEWBORNS EXCLUDED) recommendations from an expert panel from the French Intensive Care Society (SRLF) with the French Group for Pediatric Intensive Care Emergencies (GFRUP), the French Society of Anesthesia and Intensive Care (SFAR), the French Society of Hematology (SFH), the French Society for Hospital Hygiene (SF2H), and the French Infectious Diseases Society (SPILF)
Annals of Intensive Care (2016)