Published ahead of print on February 5, 2003, doi:10.1164/rccm.200206-554OC
© 2003 American Thoracic Society Acute Monocytic Leukemia Presenting as Acute Respiratory FailureMedical Intensive Care Unit and Hematology Department of the Saint-Louis Teaching Hospital and Paris 7 University; and Respiratory and Critical Care Department, Tenon Hospital, Assistance Publique-Hôpitaux de Paris, Paris, France Correspondence and requests for reprints should be addressed to Elie Azoulay, M.D., Ph.D., Service de Réanimation Médicale, Hôpital Saint-Louis, 1 Av Claude Vellefaux, 75010 Paris, France. E-mail: elie.azoulay{at}sls.ap-hop-paris.fr
Acute respiratory failure revealing acute monocytic leukemia is rare. We report 20 patients admitted to the intensive care unit (ICU) with three remarkable features: (1) rapidly progressive respiratory distress revealing acute leukemia, (2) monocytic leukemia, and (3) respiratory status deterioration after chemotherapy initiation. The median age was 50 years (17–72 years), and respiratory symptoms started 2 days (0–15 days) before ICU admission. The median leukocyte count was 98,250/mm3 (800–529,000), with circulating monocytic cells in all of the patients but one. Bone marrow examination was diagnostic of monocytic leukemia in all patients. At presentation, respiratory rate was 33 (18–50) per minute, and PaO2 on room air was 44.5 mm Hg (30–60). Chest radiographs revealed unilateral alveolar infiltrates (n = 1), bilateral alveolar infiltrates with (n = 3) or without (n = 11) pleural effusion, or diffuse interstitial infiltrates (n = 5). Alveolar hemorrhage was the main bronchoalveolar lavage finding, with monocytic cells retrieved from four patients. Respiratory function deteriorated after cancer chemotherapy initiation in all patients. Of the 15 patients who required mechanical ventilation, 10 died. Leukemic pulmonary infiltration as the first manifestation of acute monocytic leukemia should be recognized, and intensive management should be provided in anticipation of the respiratory function deterioration seen consistently after chemotherapy initiation.
Key Words: acute respiratory failure • monocytic leukemia • bronchoalveolar lavage fluid analysis • intensive care unit • alveolar hemorrhage Symptomatic leukemic infiltration of the lung is the least common cause of pulmonary infiltrates in patients with acute leukemia (1). Pneumonia (bacterial, viral, fungal, or parasitic) (2, 3) and noninfectious pulmonary abnormalities (such as edema and alveolar hemorrhage) (4–7) are both common and serious (4, 8). In patients with known acute leukemia, several elements provide diagnostic orientation, including the blast cell type, time from leukemia onset to pulmonary involvement, prior and ongoing therapies, and whether bone marrow failure is present (neutropenia and thrombocytopenia) (9). However, the definitive diagnosis depends on extensive investigations, including fiberoptic bronchoscopy with bronchoalveolar lavage (BAL) (10, 11). Surgical lung biopsy may contribute to the diagnosis (12). Both pulmonary leukostasis and leukemic infiltration of the lung have been described in patients with acute leukemia. Pulmonary leukostasis associated with lymphocytic or nonlymphocytic leukemia usually occurs in patients with leukocyte counts that are greater than 100,000/mm3 or that increase rapidly (13–15). Acute lysis pneumopathy can develop at chemotherapy initiation in these patients (16–18). In contrast, leukemic infiltration of the lungs has been reported in acute leukemia patients without hyperleukocytosis, suggesting that the type of blasts and their affinity for the pulmonary endothelium may be involved in the development of acute respiratory distress syndrome (19). We report 20 patients with acute respiratory failure from leukemic pulmonary infiltration or leukostasis caused by acute leukemia. These cases deserve special attention for three reasons: respiratory impairment was the presenting manifestation of undiagnosed acute monocytic leukemia; all patients had myeloid leukemia of the AML5 subtype; and respiratory function deteriorated in all 20 patients within a few hours after chemotherapy initiation, suggesting a need for intensive care unit (ICU) admission before starting chemotherapy.
We studied all patients with acute monocytic leukemia (AML5) presenting with acute respiratory failure and those who were admitted between January 1, 1994, and July 1, 2002, to the Medical ICU of the Saint-Louis Hospital, a 630-bed university and tertiary hospital in Paris, France. The hospital comprises 230 hematology beds, including a unit that manages only patients with acute leukemia. The medical ICU is a closed unit that admits 500 to 600 patients per year, including 20% with hematologic disorders. We reviewed ICU medical records to identify patients with acute respiratory failure revealing acute monocytic leukemia, defined as follows: acute respiratory failure antedating the diagnosis of leukemia, radiographic infiltrates, the absence of other causes of pulmonary involvement, and the absence of documented heart failure (normal echocardiography or pulmonary wedge pressure of less than 18 mm Hg). The following information was abstracted from the medical chart of each study patient: (1) epidemiologic and clinical data at ICU admission; (2) characteristics of the leukemia, including spleen, liver, or lymph node enlargement, tonsil and gingival hypertrophy, skin eruption, leukocyte count at diagnosis, percentage of circulating blast cells, hemoglobin level, platelet count, and whether disseminated intravascular coagulopathy was present; (3) at ICU admission, temperature, respiratory rate, pulmonary auscultation, and arterial blood gases on room air; (4) whether noninvasive or conventional mechanical ventilation was required in the ICU (20); and (5) results of BAL or protected specimen brush studies. Patients who spent less than 24 hours on a ward or in the emergency department before ICU admission were classified as admitted directly to the ICU. In all of the patients who met these criteria, the diagnosis of leukemia was based on cytologic examination of blood and marrow smears as previously described (19). Leukemic pulmonary involvement was classified as leukostasis if the peripheral leukocyte count was greater than 50,000 and as leukemic pulmonary infiltration if not (19). Fibroscopy with BAL was performed as previously described (see online supplement). All patients received empirical antibiotics consisting of a betalactam and a macrolide or fluoroquinolone (21, 22). Antibiotic treatment was adapted as indicated by the results of distal bronchial specimen cultures.
Statistical Analysis
Figure 1 is the flow chart of all patients with acute leukemia who were admitted to the Saint-Louis Hospital between January 1994 and July 2002. Twenty patients (13 men and 7 women) were admitted to the ICU for acute respiratory failure. Their main characteristics are reported in Table 1 . They had a median age of 50 years (range, 17–72). Their respiratory symptoms started 2 (range, 0–15) days before ICU admission. Ten patients were admitted directly to the ICU, and 10 were admitted after 2 days (range, 1–10) on a ward. At clinical examination, the respiratory rate was 33 per minute (range, 18–50). Auscultation of the lungs found crackles in all but four patients. All of the patients but two reported fever before ICU admission. At admission, the median temperature was 38.4°C (range, 36.6–40°C), but six patients had a temperature that was below 38°C. Manifestations attributable to leukemia started 15 days (range, 2–90) before hospitalization and consisted of hepatosplenomegaly or lymphadenopathy (n = 14), gingival hypertrophy (n = 10), tonsillitis (n = 7), headaches (n = 4), or a skin eruption (n = 4). PaO2 on room air at admission was 44.5 mm Hg (range, 30–60). Chest roentgenograms revealed unilateral alveolar infiltrates (n = 1), bilateral alveolar infiltrates with (n = 3) or without (n = 11) pleural effusion, or diffuse interstitial infiltrates (n = 5). Pleurocentesis was performed in two of the three patients with pleural effusions and revealed pleural exudates without monocytic cells. No clinical differences (temperature, arterial blood gazes, mechanical ventilation, and outcome) were found between patients with alveolar versus interstitial patterns by chest roentgenography.
Table 2 reports the main clinical and hematologic parameters. At admission, the leukocyte count was 98,250/mm3 (range, 800–529,000), and circulating monocytic cells (see Figure E1 in the online supplement) were seen in all the patients but one. Six patients had fewer than 20,000 leukocytes per mm3, including one patient with leukopenia (800 leukocytes per mm3). Four patients presented with disseminated intravascular coagulopathy. In all patients, bone marrow studies were diagnostic of acute monocytic leukemia, with massive bone marrow infiltration on the smears.
Five patients had severe hypoxemia precluding bronchoscopy; they underwent protected distal bronchial sampling without visual guidance. Among the 15 other patients, 1 had 102 cfu/ml of Escherichia coli, and 2 had 102 cfu/ml of Haemophilus influenzae in the BAL specimens. Findings were negative from all studies for viruses, parasites, fungi, and mycobacteria. All distal bronchial samples were negative. BAL fluid was hemorrhagic in 7 of the 15 patients in whom BAL was performed. The diagnosis of leukemic pulmonary infiltration as the cause of acute respiratory failure was based on the absence of other detectable causes in these patients with monocytic leukemia. However, in four patients, monoblastic cells were recovered from BAL fluid analysis, including one patient with alveolar hemorrhage (see Figure E2 in the online supplement). Chemotherapy was started on the day of diagnosis because of the severe hypoxemia and rapidly progressing lung infiltrates. In all 20 patients, respiratory function deteriorated within a few hours after chemotherapy initiation. Four patients experienced cardiac arrest, one patient atrial fibrillation, and one patient ventricular tachycardia. The patients who were already receiving invasive mechanical ventilation required higher FIO2 and positive end-expiratory pressure levels (n = 3), and the other patients required increased O2 levels (n = 4) or noninvasive (n = 3) and/or invasive mechanical ventilation (n = 4). No new events occurred (hemorrhage, infection, or heart failure) that could have explained the respiratory function deterioration. All patients received empirical antibiotics consisting of a betalactam and a macrolide or fluoroquinolone. Of the 15 patients who required mechanical ventilation, 3 responded adequately to noninvasive mechanical ventilation. Four required conventional mechanical ventilation after a period of noninvasive mechanical ventilation, and six required conventional mechanical ventilation from the outset. The median ICU stay was 8 days (range, 3–34). Ten patients died in the ICU from adult respiratory distress syndrome; these 10 patients were among the 15 who required mechanical ventilation. Autopsy, done in four patients, showed enlarged infiltrated lungs with diffuse interstitial and alveolar septal thickening caused by leukemic cells and diffuse intrapulmonary bleeding (see Figure E3 in the online supplement). No infectious agents were found.
We report 20 patients with acute respiratory failure related to leukemic pulmonary involvement from leukostasis or leukemic infiltration. The originality of these cases lies in the three features: (1) all patients had acute respiratory failure as the presenting manifestation of acute leukemia; (2) all patients had the same type of acute myeloid leukemia, involving monocytic cells (AML5); and (3) all patients experienced deterioration of their respiratory status after chemotherapy initiation. The diagnosis of leukemic pulmonary infiltration as the cause of acute respiratory failure is based on histologic or cytologic studies and on negative findings from a comprehensive investigation for more common causes (10, 23–25). Roentgenographic changes are nonspecific and heterogeneous, as shown by our series and by other authors (11). Rossi and colleagues reported that retrieval of leukemic cells by BAL could establish the diagnosis (26). However, only three of our patients had monoblasts in their BAL fluid. Moreover, in patients with alveolar hemorrhage, the presence of blast cells in BAL fluid is difficult to interpret. Similarly, a blast cell count above 40% in peripheral blood has been suggested as a criterion for suspecting leukemic pulmonary infiltration (11). However, four of our patients had peripheral blast cell counts below 40%, including one with 0% and one with 5%, suggesting that this criterion may be of limited diagnostic value and that other factors, such as the affinity of neoplastic cells for the pulmonary endothelium, may contribute to the development of acute respiratory distress in patients with acute leukemia (19). The index of suspicion for leukemic pulmonary infiltration should be particularly high in patients with monocytic leukemia, especially when the peripheral blast cell count is high or increases rapidly (11, 15, 27). Retrospective confirmation can be provided by the effectiveness of chemotherapy in improving respiratory function after the initial deterioration. We report a rarely described entity of leukemic pulmonary involvement in patients whose leukemia was diagnosed only after evaluation of acute respiratory failure. Although commonly found at autopsy, leukemic pulmonary infiltration is frequently suspected but infrequently documented in clinical practice (1, 9, 28, 29). Moreover, few cases of pulmonary leukemic infiltration have been published, and all occurred in patients who were known to have leukemia (11, 15, 16, 19, 26, 30, 31). Geller reported a case of acute leukemia presenting as respiratory distress in a patient with leukemic pulmonary infiltration diagnosed at autopsy (29). All of our patients presented with acute respiratory failure revealing acute leukemia, and six patients had peripheral leukocyte counts that were less than 50,000/mm3. Monocytic blasts were present in all our patients. Acute monocytic leukemia is associated with a high risk of leukemic pulmonary infiltration (32–34). In previous publications on pulmonary leukemic infiltration or leukostasis, the cell type of the leukemia varied across patients (11, 15, 16, 19, 26, 30, 31). Interactions between leukemic cells and the endothelium have been suggested as a possible mechanism for pulmonary leukostasis without hyperleukocytosis (19), and although traditionally related to "overcrowding" of leukemic blasts in the capillaries, pulmonary infiltration probably results also from direct endothelial cell damage (34). This hypothesis has received support from an experimental rat model of acute leukemia in which aggregates consisting almost solely of leukemic cells were found in medium-sized blood vessels; the aggregates completely obstructed the lung vasculature, including the largest arteries and veins, giving rise to extensive hemorrhage and edema (35). The third striking finding from our study is that all of our patients experienced respiratory status deterioration after initiation of chemotherapy, consistent with acute lysis pneumopathy. Dombret and colleagues reported two cases of hyperleukocytic myelomonocytic leukemia in which pulmonary injury was attributed to lysis of resident leukemic cells 1 and 3 days after chemotherapy initiation, respectively (16). Previous reports have acknowledged the association between chemotherapy and respiratory status deterioration in acute leukemia patients with hyperleukocytosis; they stressed the need for evaluating preinduction leukapheresis (17, 18, 36). However, half of our patients had fewer than 50,000 leukocytes per mm3, suggesting that therapeutic or even preventive strategies other than preinduction leukapheresis should be considered. Prakash and colleagues reported three patients with respiratory failure as a first or early manifestation of acute leukemia. Aggressive support with mechanical ventilation provided time to initiate appropriate chemotherapy, and one patient survived (31). Our patients received invasive or noninvasive mechanical ventilation, as needed, and immediate chemotherapy, in compliance with recommendations that chemotherapy be started without delay (13, 34). The 50% survival rate in our study provides support for this strategy. Moreover, the acute lysis pneumopathy seen in all our patients shortly after chemotherapy initiation indicates that the lungs should be evaluated routinely at the diagnosis of acute monocytic leukemia and strongly suggests that even patients with mild or moderate pulmonary infiltration should be admitted to the ICU for monitoring or treatment at chemotherapy initiation (17, 36, 37). The prognosis of critically ill cancer patients has improved over recent years (38–41). Intensive care clearly benefits cancer patients requiring noninvasive mechanical ventilation (20, 42–44). The onset of acute respiratory failure in a patient with acute monocytic leukemia should not be viewed as a terminal event but, rather, should lead to early invasive diagnostic and therapeutic management (15). In patients with any degree of respiratory impairment before chemotherapy initiation, immediate ICU admission may be warranted in anticipation of the respiratory function deterioration consistently associated with chemotherapy initiation. Advances are needed to improve our understanding of the pathophysiologic process and biologic mechanisms leading to leukostasis and leukemic infiltration.
This article has an online supplement, which is accessible from this issue's table of contents online at www.atsjournals.org Received in original form June 13, 2002; accepted in final form February 3, 2003
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
