Chapter 71. Acute Leukemia

• Many patients, especially children and younger adults, can be cured of acute leukemia with the use of intensive chemotherapy and skillful supportive care.
• The prognosis for the leukemia patient depends on the specific type of leukemia, defined by cytogenetics and immunophenotype, its stage, prior therapy, and coexisting chronic medical problems.
• Infection and bleeding are due to the severe pancytopenia caused by either the leukemia itself or its myelosuppressive treatment; these problems resolve as the normal bone marrow recovers.
• The success of chemotherapy for acute leukemia depends not only on the drug susceptibility of the individual patient's leukemia but also on the ability of that patient to survive the toxicity of treatment.
• Blood transfusions should be used to maintain a hematocrit greater than 30% and a platelet count greater than 10,000/μL.
• Hyperleukocytosis with malignant myeloblasts must be treated as a medical emergency with leukapheresis and chemotherapy.
• Central nervous system (CNS) leukemia can be treated successfully with radiotherapy, intrathecal chemotherapy, and high-dose systemic chemotherapy.
• Tumor cell lysis, aminoglycoside antibiotics, disseminated intravascular coagulation, elevated lysozyme levels, hyperuricemia, and direct infiltration by leukemia cells can lead to acute renal failure, which is often reversible with proper care.
• Necrotizing enterocolitis requires aggressive medical management with antibiotics, blood transfusions, nasogastric suction, bowel rest, and maintenance of normal serum proteins and electrolytes.
• Hyperuricemia is proportional in severity to the leukemia tumor burden, and it responds to effective chemotherapy together with administration of fluids, bicarbonate, rasburicase, and allopurinol.

Acute leukemia is a malignant proliferation of bone marrow or lymphoid cells that is uniformly fatal when left untreated. However, with the use of intensive chemotherapy and skillful supportive care, many patients, particularly children and younger adults, can be cured of this devastating disease. Not uncommonly, use of the medical ICU is necessary during the treatment of acute leukemia to manage either the complications of the disease itself or the complications of intensive cytotoxic therapy (Table 71-1). Such newly diagnosed patients deserve maximum aggressive supportive care in the medical ICU because most of the acute complications of leukemia resolve as a complete remission (CR) is achieved. The successful treatment of patients with acute leukemia requires a multidisciplinary approach with close collaboration among intensive care physicians, hematologic oncologists, infectious disease specialists, and the blood bank. Twenty-four-hour physician coverage is necessary to respond to emergencies that otherwise could lead to rapid deterioration of a patient in the midst of treatment.

Appropriate decision making requires an understanding of the pathophysiology and clinical history of the leukemic disorders. The likelihood of a successful outcome from intensive care management depends in part on the specific type of leukemia and its stage; i.e., the prognosis depends on whether the patient is newly diagnosed and not yet treated or is in an advanced stage following multiple relapses. Knowledge of prior therapy aids in the evaluation of complications, especially with regard to opportunistic infections or organ dysfunction. Since accurate diagnosis and optimal initial treatment and supportive care are essential to achieving the best chance of curing acute leukemia, it is highly desirable that newly diagnosed patients be referred to regional oncology centers where appropriate expertise and resources are readily available.

This chapter briefly reviews the treatment strategy for acute leukemia and then focuses on the serious metabolic, circulatory, and infiltrative complications that most often necessitate ICU care. The far more common bleeding and infectious complications and direct drug toxicities are dealt with in Chaps. 47, 68, 69, and 74.

Leukemia and other cancers result from genetic alterations in somatic cells. When mutational damage alters normal growth-regulatory mechanisms in a hematopoietic stem cell, a clone of neoplastic cells may proliferate, overwhelming the normal bone marrow through a selective growth advantage, although not necessarily by more rapid growth. Cooperating mutations also affect differentiation and apoptotic pathways. Leukemia cells usually retain characteristics of their original cell lineage, allowing classification1,2 (Table 71-2). However, because these cells are neoplastic, aberrant expression of proteins not usually manifested by normal cells in a particular hematopoietic lineage may be detected, yielding a “mixed cell” phenotype. The acute leukemias include acute myeloid (nonlymphocytic) leukemia (AML) and acute lymphoblastic leukemia (ALL), as well as the blastic transformations of chronic myelogenous leukemia (CML) and the myelodysplastic syndromes (MDS). Therapy-related AML occurs in patients following treatment with cytotoxic chemotherapy or radiotherapy for an earlier cancer or other disease.

Acute leukemia almost always causes marked abnormalities in the peripheral blood counts. Anemia may be mild, moderate, or severe, and the red blood cells may be macrocytic. Most patients will have moderate to severe thrombocytopenia. Although an elevated white blood cell count and readily recognizable circulating blast cells frequently are present, not all patients have leukocytosis. Some patients actually present with leukopenia, and malignant cells may not be observed in the peripheral blood. A careful review of a blood smear often demonstrates dysplasia in one or more cell lines. In elderly patients, AML frequently evolves from an MDS; mature granulocytes (called pseudo-Pelger-Huet cells) often are hypocellular with poorly condensed nuclear chromatin, and platelets may be large and may lack normal granulation. An overt or subclinical preleukemic phase may have been present for several months. The presence of greater than 20% blast cells in the marrow is the somewhat arbitrary dividing line between acute leukemia and the “preleukemic” disorder MDS.

The diagnosis and subclassification of acute leukemia rest on the interpretation of an adequate bone marrow sample. A marrow aspiration specimen, together with a bone marrow needle core biopsy from the posterior iliac crest, provides essential information. Auer rods are observed often in malignant myeloblasts. Cytochemical reactivity reveals myeloperoxidase or nonspecific esterase activity in myeloid leukemias, whereas such activity is absent in lymphoid leukemias. Immunophenotypic analysis using flow cytometry or immunohistochemical techniques detects cell-surface antigens, whose presence supports the diagnosis. Cytogenetic analysis of bone marrow specimens provides important diagnostic and prognostic information.3,4 Metaphase cells are karyotyped after direct preparation or short-term (24- to 48-hour) unstimulated culturing. Characteristic recurring chromosomal abnormalities include the presence of extra chromosomes, the loss of certain chromosomes, and structural rearrangements such as translocations, deletions, insertions, and inversions. Many distinct subtypes of acute leukemia can now be defined according to their characteristic morphologic, cytochemical, immunophenotypic, and cytogenetic features. These disorders have a more predictable clinical course than less well-defined leukemias, and this distinction is reflected in the current World Health Organization (WHO) classifications.2

The optimal chemotherapy for patients with acute leukemia has not yet been determined.5–7 Recent clinical trials have identified different risk groups that are more or less likely to respond to standard chemotherapy programs (Table 71-3). Treatment is given in two or three phases: remission induction, consolidation, and in some cases, maintenance. The goal of remission-induction chemotherapy is to induce a complete remission. Any lesser response does not improve survival. Complete remission is defined by the absence of detectable leukemia in the bone marrow and blood, the regeneration of normal marrow elements, and the return to normal levels of all three blood cell lines for at least a month. This response reflects a reduction in the tumor cell mass by more than 3 logs (i.e., >99.9%). However, as many as 108 malignant cells may still persist at this point. If no postremission therapy were given, most patients with acute leukemia would relapse within 2 to 4 months. Thus consolidation chemotherapy (sometimes called intensification therapy) is directed at destroying any clonogenic leukemia cells that remain undetected after initial treatment. Consolidation therapy uses several monthly cycles (commonly one to four) of myelosuppressive chemotherapy, each sufficiently intense to produce bone marrow hypoplasia and peripheral blood pancytopenia. Various strategies using allogeneic bone marrow transplantation or high-dose chemotherapy followed by reinfusion of autologous marrow or blood stem cells collected earlier during remission are being tested as postremission consolidation treatments. Low-dose outpatient maintenance chemotherapy that does not lower the blood count has not been proved to increase the fraction of patients with AML who are cured, although 18 to 24 months of maintenance treatment seems useful for patients with ALL.

Patients who suffer a relapse of acute leukemia after completion of all therapy often can achieve a second remission with reinstitution of the original induction chemotherapy regimen. However, patients who relapse in the midst of initial treatment or while receiving maintenance therapy are less likely to achieve a second remission and rarely have long survival. Second and subsequent periods of remission usually are progressively shorter. The risks and costs of marrow/stem cell transplantation are easily justified for patients who have had a relapse of leukemia because otherwise disease-free survival is generally (but not always) short. Patients who are younger than 50 years of age and have a human leukocyte antigen (HLA)–identical sibling are good candidates for allogeneic transplantation. Many centers offer transplantation to selected patients up to 60 years of age. It may be possible to extend the upper age limit with the use of T-cell depletion from the donor marrow to diminish the incidence and severity of graft-versus-host disease (GVHD), as well as the introduction of less toxic “reduced intensity” stem cell transplant regimens. The use of marrow stem cells from a matched but unrelated donor is also feasible.

The success of chemotherapy for acute leukemia depends not only on the drug susceptibility of the individual patient's leukemia but also on the ability of that patient to survive the rigors of treatment. Significant advances have been made in the supportive care of patients undergoing intensive cytotoxic therapy. Prevention and early treatment of infection have been foremost. Prior to treatment, all patients should have a thorough dental evaluation and a computed tomographic (CT) scan of the sinuses to identify and eliminate potential sources for infections. Patients are placed in protective isolation in single rooms; in this way, these immunocompromised patients are protected from common infectious agents in the hospital environment. Laminar airflow rooms are not necessary for successful treatment, but patients generally are restricted from eating fresh fruits and vegetables to decrease their enteric bacterial flora. Strict attention to thorough handwashing by all medical and nursing personnel and visitors is necessary.

Patients with leukemia are severely immunocompromised by both their disease and their treatment. Because chemotherapy damages mucosal barriers, these patients are susceptible to infection by endogenous organisms. Most common are gram-negative enteric bacteria, gram-positive cocci, and fungi such as Candida and Aspergillus species. Patients with ALL also are susceptible to Pneumocystis, mycobacterial, and viral infections. Prophylactic use of oral antibiotics such as fluoroquinolones can decrease the incidence and severity of infections in leukopenic patients. Antiviral and/or antifungal prophylaxis is also employed commonly in patients with profound or prolonged leukopenia.

Standard practice dictates that broad-spectrum antibiotic therapy must be initiated promptly and empirically in a granulocytopenic patient at the time of the first fever greater than 38.5°C (101.5°F). The Infectious Disease Society of America recently published guidelines on the management of these patients.8 Blood, urine, and sputum cultures should be obtained, but the source of infection is rarely identified in part because there is little inflammatory response in the tissues. The first priority must be to prevent septic shock. Ceftazidime or imipenem as a single agent or the combination of a semisynthetic penicillin plus an aminoglycoside is widely used, but the choice of regimen depends in part on renal function and history of allergies. High-risk patients also should be given vancomycin without delay because nosocomial gram positive organisms may be resistant to methicillin. Should initial cultures reveal a pathogen, antibiotic therapy should be adjusted for maximum bactericidal activity based on susceptibility testing in vitro.

Pulmonary infiltrates in febrile leukemia patients should be managed aggressively in an attempt to identify the etiology and treat the underlying cause (Table 71-4). In general, however, broad-spectrum treatment continues until the chemotherapy is completed and the granulocyte count recovers to near normal.

Patients with acute leukemia are characterized by bone marrow failure owing either to replacement by leukemia cells or to chemotherapy-induced hypoplasia. Red blood cell transfusions are given to maintain a hematocrit of approximately 30%. A higher hematocrit should be maintained in hypoalbuminemic patients to keep the blood volume expanded and tissues perfused. While the general practice in critically ill patients has been to restrict red blood cell transfusion until the hematocrit falls much lower than this threshold, we believe that leukemia patients are a unique population not well represented in trials of transfusion strategies. If red blood cell transfusion is, in fact, immunosuppressive, it is unlikely to have much additional impact in chemotherapy-immunosuppressed leukemic patients. Further, these patients may benefit from the additional margin of safety that a higher hematocrit provides in case of thrombocytopenia-related hemorrhage.

Bleeding is a common problem, and the risk rises as the platelet count decreases (see Chap. 69). Platelets should be transfused prophylactically to maintain a platelet count of greater than 10,000/μ L to decrease the risk of spontaneous hemorrhage, such as a stroke or gastrointestinal bleeding. A higher platelet count is necessary to treat active bleeding. No salicylates or other drugs that interfere with platelet function should be given, nor should intramuscular injections be given to a thrombocytopenic patient. The use of single-donor platelets collected by apheresis will decrease the rate of alloimmunization in leukemia patients. Patients who become alloimmunized may benefit from HLA-matched platelets or antibody–cross-matched platelets to achieve adequate posttransfusion platelet counts. Blood bank techniques to remove leukocytes by filtration or centrifugation from red blood cell or platelet transfusion units may slow the rate of alloimmunization.

Granulocyte transfusions are rarely necessary for patients receiving chemotherapy for acute leukemia. The indications for their use generally are limited to severely granulocytopenic patients who remain febrile and bacteremic despite receiving antibiotics that have bactericidal activity against the particular organism in vitro. In this situation, the phagocytic activity of exogenous granulocytes can be lifesaving. One or two leukapheresis products are transfused daily for 4 to 7 days. Complications include respiratory distress from leukoagglutination, transmission of cytomegalovirus (CMV), and rapid alloimmunization. Pulmonary infiltrates may worsen as granulocytes migrate into infected lung tissues.

Over the past several years, the hematopoietic growth factors filgrastim (granulocyte colony-stimulating factor [G-CSF]) and sargramostim (granulocyte-macrophage colony-stimulating factor [GM-CSF]) have become available commercially in the United States. Pegfilgrastim (neulasta), a long-acting form of G-CSF also has become available recently for clinical use. They are now used routinely to support cancer patients with solid tumors or lymphoma who have received myelosuppressive chemotherapy. The American Society of Clinical Oncology has adopted guidelines for the use of colony-stimulating factors in clinical practice.9 Colony-stimulating factors have been shown to reduce significantly the incidence of febrile neutropenia when used prophylactically in certain patients who receive intensive chemotherapy regimens. Specifically, it is recommended that patients who receive a chemotherapy program previously associated with a greater than 40% incidence of febrile neutropenia be treated prophylactically with colony-stimulating factors to reduce this risk. In addition, there is evidence that administration of colony-stimulating factors may reduce the risk of febrile neutropenia in subsequent cycles of chemotherapy for patients who had a documented occurrence in an earlier cycle. Most randomized, controlled trials have not demonstrated significant long-term clinical benefit (response rate or outcome) from using myeloid growth factors in patients with AML. However, modest benefits in decreased duration of neutropenia and hospitalization have been observed. For adults with ALL, clinical trials have demonstrated statistically significant shortening of the duration of neutropenia following remission-induction chemotherapy when G-CSF was used. The clinical benefit was greater in older patients.

The current data do not support the routine use of colony-stimulating factors as adjuncts to antibiotic therapy in febrile neutropenic patients. Nevertheless, their use along with antibiotics may be reasonable in high-risk patients. Specifically, neutropenic patients with clinical features that predict a poor outcome, such as pneumonia, documented fungal infection, or sepsis syndrome, are reasonable candidates for therapy with colony-stimulating factors.9

Soft Silastic right-atrial catheters (e.g., double- or triple-lumen Hickman or Groshong catheters) that tunnel under the skin and enter the innominate vein are used widely. They allow direct access to a large vein for infusion of chemotherapy agents, antibiotics, and blood transfusions and also allow painless blood sampling for laboratory monitoring. Parenteral hyperalimentation can be administered easily through these catheters, if necessary. Although a boon to both patient and physician, such chronic indwelling lines are not without hazard. They must be placed in a sterile fashion by a skilled surgeon or invasive radiologist, especially in a thrombocytopenic patient. A vigorous skin preparation preoperatively reduces the likelihood of a staphylococcal or corynebacterial subcutaneous tunnel infection or bacteremia. Although most gram-negative bacteremias can be treated successfully in a granulocytopenic patient by antibiotics alone without removing the catheter, gram-positive infections or candidal infections of the tunnel or vein may require catheter removal.

A small proportion of patients with leukemia have an extraordinary elevation of circulating leukocytes. Hyperleukocytosis (>100,000 blast cells per microliter) is a true medical emergency, but the risks of circulatory complications begin to rise at counts above about 50,000/μ L (Table 71-5). These patients present special problems because of the rheologic effects of leukemia blast cells in the circulation of the lung, brain, and other organs; the metabolic consequences when massive numbers of leukemia cells are destroyed simultaneously by cytotoxic drugs; and perhaps other mechanisms.10 Although in some cases the blood viscosity is markedly increased because of the elevated leukocrit, in most cases the whole blood viscosity will not be increased above normal because hyperleukocytosis is associated most frequently with severe anemia.11 However, red blood cell transfusions can precipitate a hyperviscosity syndrome in a patient with hyperleukocytosis and should be delayed, if possible, until the white blood cell count falls. Nevertheless, viscosity may be high in the microcirculation. The high oxygen consumption and invasiveness of leukemia cells may interact with slow flow through the capillaries to lead to hypoxemia and vascular damage.

The contribution of blood cells to the viscosity of blood is a function of the deformability of individual cells and their fractional volume. Immature leukocytes are much larger and less deformable than erythrocytes. Leukemic myeloblasts are considerably larger than lymphoblasts, which are, in turn, larger than leukemic lymphocytes. Thus the incidence of significant leukostasis is highest in CML in the blast phase, followed by AML and then ALL. In contrast, leukostasis is rare in patients with chronic lymphocytic leukemia (CLL) despite white blood cell counts as high as 500,000/μ L. Similarly, well-hydrated patients with CML in the chronic phase can tolerate total white blood cell counts of 200,000/μ L without difficulty because only a small fraction of the cells are blasts or promyelocytes, and most are mature granulocytes.

Leukostasis and hypoxia in the capillary beds can induce respiratory distress, cardiac arrhythmias, and central nervous system (CNS) symptoms leading to coma. Death follows rapidly unless the circulation can be restored. Emergency measures include leukapheresis to remove a large mass of tumor cells directly from the bloodstream.10 A single efficient leukapheresis can decrease the leukocyte count by approximately 50% within 2 to 3 hours. A centrifuge technique rather than a filtration method should be used because blast cells do not adhere to glass-wool filters. Prompt use of antimetabolite drugs such as oral hydroxyurea or intravenous cytarabine can reduce cell proliferation rapidly and decrease the blast count transiently. A single dose of cranial irradiation (200 to 400 cGy) can ameliorate CNS symptoms due to leukostasis. Diffuse pulmonary infiltrates and respiratory distress also may respond to thoracic irradiation. Once the circulation through the capillaries has been restored, systemic chemotherapy should be initiated promptly. Prophylactic platelet transfusions must be given to prevent bleeding as the circulation is restored to hypoxic tissues. The invasiveness of leukemic myeloblasts may lead to injury and disruption of vascular endothelium. Lysis of these cells in situ releases proteolytic enzymes, causing further tissue damage and bleeding.

Tachypnea and pulmonary infiltrates in a patient with hyperleukocytosis can mimic pulmonary embolism or infection. Care must be taken to rule out “pseudohypoxemia,” which results from the in vitro consumption of oxygen within an arterial blood-gas sample by blast cells in transit to the laboratory.12 Such blood specimens should be transported rapidly on ice. Rarely, marked thrombocytosis can produce a similar artifact in addition to pseudohyperkalemia.

Leukemia is, from its genesis in the marrow, spleen, or thymus, a systemic disease. Malignant cells circulate through the bloodstream and lymphatics and can infiltrate the parenchyma of any tissue. ALL cells migrate preferentially to lymphoid organs such as lymph nodes, spleen, and liver, causing organomegaly. When malignant myeloblasts form a solid tumor, a granulocytic or myeloid sarcoma results. The older term chloroma (“green cancer”) describes the greenish color that results from the myeloperoxidase reaction when a myeloid sarcoma is exposed to air. Myeloid sarcomas occur most often in patients with CML in blast phase or with myelomonocytic or monoblastic leukemia, and most commonly they involve bone and subperiosteum, lymph nodes, skin, soft tissues, and the gastrointestinal tract.13 They most frequently cause pain but also may cause neuropathy or intestinal obstruction from their mass effect. They are quite responsive to local irradiation (1000 to 2000 cGy), but since they are a localized manifestation of a disseminated disease, systemic chemotherapy also must be used. Local therapy often is not necessary because myeloid sarcomas often shrink and disappear within a few days after chemotherapy begins.

CNS infiltration by leukemia cells is more common in ALL than AML and is more common in children than in adults. Among patients with myeloid leukemia, the incidence of CNS involvement is greatest in children and in those with monoblastic leukemias or high circulating blast counts. Among the lymphoblastic leukemias, the risk is greatest in patients with the T-cell or mature B-cell (Burkitt) immunophenotypes.14 When the vascular leptomeninges surrounding the brain and spinal cord are infiltrated by malignant cells, leukemia cells often (but not always) can be found in the cerebrospinal fluid (CSF) by lumbar puncture. Cranial neuropathy and spinal radiculopathy result from the impingement of an expanding mass of leukemia cells on these nerves where they traverse narrow bony foramina. In contrast, a few malignant cells in the CSF at diagnosis probably are not significant because they often disappear during the initial induction chemotherapy treatment without specific attention to the CNS.

The most frequent symptoms in patients with overt CNS leukemia are those caused by increased intracranial pressure: vomiting, headache, papilledema, and lethargy. Diplopia results from oculomotor nerve palsy, especially of the abducens nerve. Facial nerve palsies are also common. Spinal radiculopathy and the cauda equina syndrome occur more commonly in adults than in children. Mass lesions in the brain substance itself or in the spinal cord are not common, and seizures are rarely present. Meningismus is also uncommon, and its presence suggests infection or subarachnoid hemorrhage.

The diagnosis is established by finding leukemia cells in the CSF. Usually, the CSF pressure is elevated in symptomatic patients, the glucose concentration is low, and the protein concentration is moderately elevated. CT scan or magnetic resonance imaging (MRI) may demonstrate thickening or contrast enhancement of the membranes around the base of the brain or the spinal nerves. Because spinal subarachnoid hemorrhage or focal hematomas can follow lumbar puncture in thrombocytopenic patients, platelets should be transfused to counts of greater than 20,000/μ L and any coagulopathy corrected before a careful lumbar puncture is done.

CNS leukemia is usually rapidly responsive to irradiation or intrathecal chemotherapy. Standard treatment for symptoms is 2400 cGy to the whole brain in 12 fractions. The posterior orbits and upper cervical spine should be included. The spinal column is rarely irradiated except in patients with symptomatic radiculopathy because of the resulting damage to the underlying bone marrow. Dexamethasone 4 mg four times per day can alleviate symptoms of increased intracranial pressure. The spinal cord is best treated with preservative-free methotrexate (12 mg/m2 to a maximum dose of 15 mg) injected intrathecally every 3 days for at least six doses. Once the CSF is clear of malignant cells, intrathecal therapy should be administered once a month for the following year. Methotrexate can cause an inflammatory arachnoiditis with fever and nuchal rigidity. For this reason, hydrocortisone (50 mg) often is coadministered with the methotrexate. Alternative drugs for intrathecal use are cytarabine (DEPOCYTE,) 50 mg), or less commonly, thio-TEPA. Slow diffusion of methotrexate out of the CSF into the systemic circulation can cause myelosuppression and gastrointestinal toxicity. This problem can be lessened by giving 5 mg leucovorin 36 hours after each dose. Leucovorin decreases methotrexate's toxic effects on normal tissues by replenishing the supply of reduced folate to cells. Too much leucovorin can interfere with subsequent doses of methotrexate. Administration of intrathecal chemotherapy via lumbar puncture has the disadvantage of not treating the ventricles at all and treating the top surface of the brain only poorly. For this reason, an Ommaya reservoir often is placed in symptomatic patients who require chronic CNS therapy to allow an intraventricular (and easier) route of administration.

Clinical trials have suggested that the CNS is a sanctuary site for leukemia cells and is poorly treated by many intravenously administered drugs. Patients can have an isolated CNS relapse even while they are in an apparent bone marrow remission. To diminish this risk in patients with ALL, all patients with lymphoblastic leukemia generally receive some prophylactic treatment of the CNS as part of their initial treatment program. Alternatives include high doses of cytarabine or methotrexate, two drugs that cross the blood-brain barrier and reach cytotoxic concentrations in the CSF even when administered intravenously. CNS prophylaxis is not commonly employed for AML.

Although uncommon, various paraneoplastic syndromes occur in patients with leukemia, such as Guillain-Barré polyneuritis and progressive multifocal leukoencephalopathy. It is not known to what extent the leukemia itself or its treatment with chemotherapy or irradiation is responsible for these disorders. High doses of cytarabine (2000 mg/m2 for 6 to 12 doses) can cause irreversible cerebellar ataxia, especially in elderly patients and those with renal failure.

Marked leukemic infiltration of the liver and spleen is infrequent in AML and occurs principally in patients with ALL or CML in the blast phase. Signs and symptoms may mimic those of an acute hepatitis, with jaundice, tender hepatomegaly, and elevated serum transaminase levels. This problem usually resolves with effective chemotherapy, but the choice of initial chemotherapy may be restricted because many useful drugs are cleared through the liver. A biliary ultrasound examination should be performed to rule out bile duct obstruction. The spleen is rarely massively enlarged, although it may be palpable at diagnosis owing to leukemic infiltration. Vascular infarction can occur, causing painful subcapsular hematomas. A rub with respiratory variation sometimes will be heard over the left upper quadrant. Rarely, such an infarcted spleen ruptures, causing shock and signs of intraperitoneal hemorrhage. This is a surgical emergency. Splenic rupture also can occur during chemotherapy as the tumor cell mass is receding, but this event is extremely rare.

Infiltration of the kidneys is more common in ALL and monoblastic leukemias than in other types of AML. These patients present with oliguric acute renal failure. Renal ultrasound examination demonstrates homogeneous enlargement of both kidneys but no ureteral obstruction. Alternatively, enlarged retroperitoneal lymph nodes secondary to infiltration by leukemia cells can cause ureteral obstruction and renal failure without intrinsic renal disease. These signs resolve during the course of effective chemotherapy, but the inability to excrete the breakdown products of tumor cell lysis can precipitate hyperkalemia and hyperuricemia with their resulting metabolic complications (see below). One or two doses of radiation therapy (200 to 400 cGy) to include just the kidneys if they are enlarged (or the ureters as necessary) can re-establish renal excretory function prior to the use of systemic chemotherapy and thus alleviate the risk of tumor lysis syndrome.

Typhlitis is a necrotizing enterocolitis particularly of the terminal ileum, appendix, cecum, and right colon that occurs in granulocytopenic patients.15 Although seen most often in young patients who have recently received intensive chemotherapy, typhlitis can occur in any granulocytopenic patient, such as those with aplastic anemia or acute leukemia not yet treated. Symptoms and signs are similar to those of inflammatory bowel disease: nausea, vomiting, abdominal pain and tenderness, profuse watery or bloody diarrhea, and fever. The intestinal mucosa is ulcerated, allowing enteric organisms to invade into and through the bowel wall. Ileus and bowel dilation result. Plasma proteins and electrolytes are lost into the bowel lumen, as occurs with toxic megacolon. Bowel perforation and peritonitis may follow. Jaundice and hepatitis are common, probably because of bacterial seeding through the portal vein.

Most patients are best managed with aggressive medical treatment: broad-spectrum antibiotics; transfusion of red blood cells, platelets, and fresh frozen plasma; maintenance of normal serum electrolyte levels (especially of potassium); and bowel rest with nasogastric suction. Narcotic analgesics and bowel paralytic agents such as LOMOTIL diphenoxylate hydrochloride and atropine sulfate) should be avoided because they increase the risk of ileus. Granulocyte transfusions in this setting can prove lifesaving in severely neutropenic patients. With appropriate treatment of the transmural gastrointestinal infection and the recovery of granulocytes, this problem usually resolves without sequelae. Patients with clear evidence of bowel perforation require surgery with ileal diversion or bowel resection. Peritoneal lavage can establish the correct diagnosis if blood or feculent material is recovered. Although the perioperative mortality is high in such cases, the likelihood of a granulocytopenic patient surviving a disseminated peritonitis without surgery is small. Radiographic evidence of air within the bowel wall (pneumatosis intestinalis) does not in itself necessitate surgery, nor should a granulocytopenic patient who is otherwise doing well necessarily undergo laparotomy for intraperitoneal free air. Antibiotic-associated pseudomembranous colitis (Clostridium difficile), radiation enteritis, and graft-versus-host disease (GVHD) are other causes of abdominal pain and bloody diarrhea in leukemia and stem cell transplant patients.

Uric acid is the product of nucleic acid catabolism, and its level is increased in diseases in which cell turnover rates are high. Uric acid is excreted largely by the kidneys through both glomerular filtration and tubular excretion. Uric acid nephropathy in leukemia may be present at diagnosis, or it may occur as a complication of therapy. Hyperuricemia may be exacerbated and uricosuria increased as a result of successful cytolytic therapy, as well as by drugs such as thiazide diuretics, probenecid, and aspirin. Hyperphosphatemia due to tumor cell lysis can cause deposition of calcium phosphate precipitates in the calyces and tubules. In acid urine, uric acid quickly supersaturates and precipitates. Uric acid has a dissociation constant of 5.4. At pH 7.0 and above, over 99% of urates are in the ionized form and thus are water-soluble. At pH 4.5, on the other hand, 80% of urate is in the form of poorly soluble uric acid, and the combined solubility of uric acid and urate falls below 8 mg/dL. In leukemia patients, therefore, the production of organic acids, together with the decreased urine formation that results from dehydration owing to anorexia or fever, frequently leads to precipitation of urates in the renal tubules and collecting system. Direct infiltration of leukemia cells into the kidneys also appears to enhance the local accumulation of uric acid. Leukemia patients with prior renal disease are more likely to develop uric acid nephropathy. Intratubular precipitation of urates leading to obstruction of urine flow is the more critical cause of nephropathy rather than renal cell injury per se.

Effective management of hyperuricemia and hyperuricosuria in leukemia patients is based on reducing the production of uric acid and at the same time promoting the solubility of uric acid in the urine. An adequate urine flow (100 mL/h) should be established by oral and intravenous hydration. The blood volume must be expanded and acidosis corrected. Acetazolamide, a carbonic anhydrase inhibitor, can be given together with sodium bicarbonate to alkalinize the urine. The optimal urine pH is 7.0 to 7.5. Allopurinol (300 mg/d) will effectively inhibit the conversion of xanthine and hypoxanthine to uric acid. Allopurinol treatment with alkalinization of the urine markedly decreases the risk of uric acid nephropathy in nearly all cases of leukemia. Rasburicase (ELITEK) (0.20 mg/kg) is a recombinant human urate oxidase that, when administered intravenously, can rapidly metabolize serum uric acid. It may be used daily for 1 to 5 days when rapid decreases in serum uric acid are needed or when a large tumor cell lysis is predicted.16 Allopurinol should not be used in combination with rasburicase but can be restarted after the hyperuricemia resolves. Rasburicase is not recommended for patients with glucose-6-phosphate dehydrogenase (G6PD) deficiency. Once the tumor cell mass has been reduced, these measures can be discontinued safely. Allopurinol causes an allergic dermatitis in approximately 10% of patients, but this side effect usually occurs only after more than a week of treatment, allowing time for leukemia cytoreduction. Fever and the Stevens-Johnson syndrome may follow the systemic rash if allopurinol is continued.

Lysozyme is a small protein (molecular weight = 14,000) with an unusually high isoelectric point (pH = 11) that has a remarkable capacity to lyse the cell walls of certain bacteria. Lysozyme (also known as muramidase) is present in high concentrations in granulocytes and monocytes as well as tissue macrophages. Patients with monocytic and myelomonocytic leukemia excrete large quantities of lysozyme in the urine.17 Lysozyme is toxic to renal tubule cells and causes tubular dysfunction. Marked hypokalemia generally has accompanied lysozymuria in patients with monoblastic and myelomonocytic leukemias. These problems resolve with effective treatment of the leukemia.

Laboratory evidence of disseminated intravascular coagulation (DIC) is almost always present at diagnosis in patients with acute promyelocytic leukemia (APL), but it can accompany any leukemia or infection where there is rapid cell lysis or tissue destruction. The pathophysiology is still controversial, but the coagulation cascade appears to be initiated by the release of intracellular thromboplastic material.18 The syndrome is characterized by the consumption of plasma coagulation proteins and platelets, resulting in prolongation of the prothrombin time (PT), thrombin time (TT), and partial thromboplastin time (PTT), together with a decrease in fibrinogen levels and an elevation in the levels of fibrin degradation products (FDPs) and D-dimer. Oozing from venipuncture sites is common. Fibrin thrombi usually are present in the renal glomeruli, leading to moderate renal insufficiency. Aminoglycoside antibiotics should be avoided because patients with DIC are unexpectedly susceptible to renal failure despite blood level monitoring. Trauma to circulating red blood cells from fibrin strands creates schistocytes, which are apparent on the blood smear. Deep vein thrombosis and pulmonary embolism may occur. The DIC in patients with leukemia is most intense shortly after chemotherapy begins, when the destruction of leukemia cells results in the release of procoagulant-rich granules. The DIC wanes as the tumor burden declines. However, a state of low-grade DIC may persist in the setting of ongoing tumor lysis, bacteremia, or sepsis.

Management of DIC is always directed at correcting the underlying tissue destruction, whether due to leukemia, infection, a burn or crush injury, or an obstetric complication. In the meantime, the consumption of coagulation proteins through fibrin formation and its associated secondary fibrinolysis can be brought under control by administration of exogenous cryoprecipitate, platelets, and fresh frozen plasma, as well as by the use of heparin. A continuous intravenous infusion of heparin at 5 units/kg per hour is well tolerated even in thrombocytopenic patients and allows fresh frozen plasma or cryoprecipitate to be transfused safely without concern about adding substrate for renewed intravascular coagulation. Nonetheless, there are no randomized clinical trials demonstrating benefit from the use of heparin products in DIC. Subcutaneous and/or low-dose intravenous heparin should be used with great care in patients with DIC and particularly in actively bleeding patients. Cryoprecipitate infusions should be used to maintain the fibrinogen concentration at greater than 100 mg/dL. Several doses of vitamin K (10 mg/d) should be given. If the liver is functioning normally, the vitamin K–dependent factors II, VII, IX, and X, as well as antithrombin, will be repleted rapidly, thereby avoiding the need for transfusion of large volumes of fresh frozen plasma. If after 24 hours of heparin therapy the fibrinogen level has not stabilized and the FDP levels have not decreased, the heparin infusion should be increased to 10 U/kg per hour; a higher dose of heparin is rarely necessary. Lack of response to heparin may indicate a deficiency of the heparin cofactor antithrombin, which can be repleted by transfusion of fresh frozen plasma. The optimal approach to DIC in leukemia remains to be determined. It is usually possible to manage the coagulopathy associated with APL with all-trans-retinoic acid, intensive chemotherapy, and blood product support without the routine use of heparin.19,20

All-trans-retinoic acid (tretinoin, ATRA) has emerged as a remarkably effective agent for inducing complete remissions in patients with acute promyelocytic leukemia (AML-M3 or APL).21 Patients with this form of AML have a characteristic chromosomal translocation, t(15;17). The juxtaposition of the retinoic acid receptor-α (RAR-α) gene on chromosome 17 and the PML gene on chromosome 15 results in the expression of fusion gene products in the leukemia cells. The leukemic blasts undergo terminal differentiation in response to ATRA, and complete responses are achieved in the majority of patients. The best results are achieved when patients also receive cytotoxic chemotherapy concomitantly with ATRA. The success of ATRA has led to its incorporation into remission-induction regimens for all patients with APL.

ATRA has several advantages over standard induction chemotherapy: (1) Patients usually do not have an exacerbation of DIC with ATRA treatment, as can happen with chemotherapy. In fact, resolution of DIC often occurs within 24 to 48 hours. (2) Patients treated with ATRA alone do not suffer bone marrow aplasia and prolonged neutropenia, as do patients treated with chemotherapy. (3) The systemic side effects are generally less severe with ATRA than with chemotherapy.

However, ATRA does have its own unique toxicities, which must be anticipated to maximize the safety of this novel agent (Table 71-6). The two most significant toxicities of ATRA are hyperleukocytosis and the retinoic acid syndrome.22 In contrast to patients receiving chemotherapy who undergo a rapid cytoreductive phase accompanied by increased DIC and tumor lysis syndrome, APL patients receiving ATRA sometimes exhibit a progressive leukocytosis, which may become marked. The leukocytosis occurs because ATRA causes differentiation of the leukemic blasts in the bone marrow, resulting in the accumulation of maturing forms in the peripheral blood. This is less likely to occur if early institution of concurrent cytotoxic chemotherapy (usually daunorubicin or idarubicin) at full doses or hydroxyurea is used. Leukapheresis has also been used to reduce the white blood cell count in these patients.

The retinoic acid syndrome occurs in approximately 25% of patients receiving ATRA and is characterized by fever, rapidly progressive respiratory distress with diffuse pulmonary infiltrates on chest radiograph, pleural effusions, and weight gain (Table 71-7). If the syndrome is unrecognized and left untreated, progressive hypoxemia, multiorgan failure, and death frequently occur. The syndrome may occur as early as 2 days after the initiation of treatment or as late as 3 or 4 weeks. The syndrome may be more common in patients with ATRA-induced leukocytosis, but it also occurs in patients with white blood cell counts of less than 10,000/μ L. Clinically, the syndrome is similar to “capillary leak” syndromes associated with certain cytokines, such as interleukin 2. This fact has led to the hypothesis that the retinoic acid syndrome may itself be cytokine-mediated. Similar symptoms may occur in chemotherapy-treated APL patients, implicating the malignant cells and not the treatment as the primary problem. Early recognition of the syndrome and prompt treatment with steroids (e.g., dexamethasone 10 mg bid for 3 days) effectively reverse the syndrome in most patients. The syndrome may be confused with an infectious complication. Thus it is important to recognize the retinoic acid syndrome promptly so that patients are not treated solely for infection, which could delay the start of corticosteroid treatment.

Another potential complication of ATRA is intracranial hypertension, or “pseudotumor cerebri.” This complication is more frequent in pediatric patients receiving ATRA. Such cases of pseudotumor have been treated with lumbar puncture, corticosteroids, and narcotic analgesia.

Marked elevations of blood lactate or pyruvate concentrations in patients with leukemia result from an increase in the rate of lactic acid production or a decrease in the rate of pyruvic acid use or both. Whether clinical acidosis results depends in part on other metabolic and respiratory factors that influence acid-base balance. Blood lactate elevations occur more commonly in patients with acute leukemia than in those with any other neoplastic disorder.23 Lactic acidosis also may be secondary to inadequate oxygenation of the tissues due to hyperleukocytosis, septic shock, pulmonary edema, or severe anemia. The development of lactic acidosis despite an apparently adequate supply of oxygen to the tissues sometimes reflects disseminated fungal infection with small-vessel occlusion or extensive liver necrosis. Lactic acidosis associated with massive leukemia is seen most commonly in patients with ALL and especially mature B-cell or Burkitt ALL. Experiments in vitro suggest that although malignant leukocytes exhibit little lactic acid production under aerobic conditions, they can have considerable anaerobic lactic acid production. It has been estimated that a rapidly growing mass of leukemia cells crowded into organs where the cells may be poorly oxygenated and therefore are dependent on anaerobic glycolysis for energy could produce as much as 300 to 600 mM lactic acid per hour. Any interference with hepatic blood supply or hepatic metabolism can diminish the body's ability to handle this acid load, resulting in clinical acidosis.

The successful treatment of lactic acidosis in patients with acute leukemia depends on the institution of effective antileukemia therapy. Removal of lactic acid by hemodialysis or peritoneal dialysis also may be useful.