Leukocytapheresis for AML hyperleukocytosis: Failing to make the grade

Hyperleukocytosis, defined as a white blood cell (WBC) count of greater than 50 to 100 × 10⁹/L, is recognized as a poor prognostic feature in patients with acute myeloid leukemia (AML) and is associated with a high risk of early mortality. When AML presents with hyperleukocytosis, death within the first 30 days has been reported to occur in 17% to 57% of cases.^{1, 2} One major predictor of early death and reduced overall survival with hyperleukocytic AML is the development of leukostasis, characterized by occlusion of small vessels with high numbers of rigid AML blasts that adhere to the endothelium, cause ischemic injury, and infiltrate into the tissue. The lungs and central nervous system (CNS) are the major affected end organs, but retinal, renal, cardiac, and gastrointestinal leukostasis complications also occur. Pulmonary manifestations of leukostasis include hypoxia, diffuse alveolar hemorrhage, and respiratory failure, while CNS complications include confusion, headache, intraparenchymal bleeding, and coma. Other significant complications associated with hyperleukocytosis include tumor lysis syndrome and disseminated intravascular coagulopathy (DIC). Hyperleukocytosis in AML patients with impending or definitive leukostasis is well recognized as a medical emergency. Novotny and colleagues have developed a clinical grading score allowing one to determine the probability of leukostasis in a patient with leukemia with hyperleukocytosis.³ However, this grading score has not been widely adopted in routine clinical practice, and diagnostic criteria for leukostatsis have not been standardized.

Interventions to rapidly cytoreduce hyperleukocytosis with AML include individual or combination treatments with hydroxyurea, induction chemotherapy, and leukocytapheresis.⁴ The cytoreductive effect of hydroxyurea is less rapid than induction chemotherapy; hydroxyurea is often given as a bridging strategy in patients where the hematologic malignancy is not yet diagnosed or where there is contraindication to induction chemotherapy. During leukocytapheresis, the patient is connected to an apheresis device and WBCs (including leukemic blast cells) are isolated from the patients' blood by density gradient separation and removed with or without the addition of replacement fluid such as colloid and/or crystalloid solution. A single procedure can reduce the WBC count by 30% to 60%⁵; however, once the procedure is completed, the WBC count will rapidly rebound if cytoreductive chemotherapy has not been started. Evidence for the clinical benefit of leukocytapheresis in patients with symptomatic leukostasis consists of observational case reports and retrospective cohort data.⁵ Mechanical reduction of the intravascular leukemic cellular burden by leukocytapheresis is thought to facilitate the reversal of leukostatic microvascular occlusions and improve tissue perfusion. Although multiple studies indicate dramatic improvements in end-organ complications, with some having observed a reduction in early mortality,^6-8 other studies report no benefit^9-11 and have raised concerns that leukocytapheresis might delay the timely start of induction chemotherapy.

The report by Bewersdorf et al in this issue of TRANSFUSION describes the largest systematic review and meta-analysis to date of AML patients receiving leukocytapheresis for the management of hyperleukocytosis, defined as a WBC count greater than 50 × 10⁹/L. A strength of this review is that it included only published studies with a control group of patients who did not receive leukocytapheresis. Concurrent cytoreductive treatments were not reported in many studies and were not included in the analyses. The primary outcome was the risk ratio of early death (up to 30 days) of those undergoing leukocytapheresis compared to those without this intervention. A secondary outcome was the odds ratio of clinical leukostasis among treated and untreated patients. The investigators included rigorous assessments of study quality and bias using the GRADE (Grading of Recommendations Assessment, Development and Evaluation) approach. Sensitivity analysis compared endpoints after excluding the study with the largest impact on heterogeneity. The authors also determined whether outcomes were affected by including studies that did or did not use a matching process, such as propensity score matching.

From among 13 two-arm retrospective studies (1743 patients), the early mortality rate in those treated with leukocytapheresis (486 patients) was not improved compared to those that did not receive leukocytapheresis (1257 patients; relative risk, 0.88; 95% confidence interval, 0.69-1.13; P = .32). Patients with clinical leukostasis were twice as likely to receive leukocytapheresis; however, this trend was not statistically significant. The authors acknowledge the inherent risks of confounding by inconsistent definitions and misclassification of leukostasis among the studies. However, their results concur with two recent comparative studies that used a more stringent propensity score matching approach to control for differences in baseline characteristics, leukostasis, and the extent of hyperleukocytosis.^{1, 11} The lack of effect in a prespecified sensitivity analysis that excluded the study with greatest heterogeneity also reinforces the conclusion of the meta-analysis that leukocytapheresis does not significantly improve early mortality.

The observations by Brewersdorf et al align with the results of a systematic review and meta-analysis by Oberoi et al¹² in 2014, which found that intervention with leukocytapheresis or the use of hydroxyurea and/or low-dose chemotherapy did not impact early mortality (7-42 days) for adult and pediatric patients with AML and hyperleukocytosis. Of note, the Oberoi study included a variety of cytoreductive strategies and cranial irradiation, making the results difficult to interpret. Some studies have included patients who went on to receive concurrent or sequential induction chemotherapy immediately after leukocytapheresis.¹¹ One may ask, might early intensive chemotherapy mask the benefit of leukocytapheresis? A recent survey of 12 international centers' experience using leukocytapheresis in older adults (>66 years old) with AML and hyperleukocytosis but without intensive chemotherapy confirmed the poor prognosis associated with clinical leukostasis. Moreover, although there was a trend toward improved outcomes, no statistically significant benefit of leukocytapheresis was observed for the 32 treated patients, compared to 187 patients who did not receive leukocytapheresis.²

There is no dispute that leukocytapheresis can achieve rapid cytoreduction. The procedure is not without complications, however. A large-bore temporary central venous catheter (CVC) is often required. CVC insertion can be challenging in the setting of thrombocytopenia and coagulopathy with DIC. Complications include bleeding, thrombosis, infection, and, less commonly, malposition. Fluid shifts during leukocytapheresis are also a concern, especially in patients with renal failure and bleeding. The use of citrate (anticoagulant citrate dextrose solution A) as anticoagulant in the apheresis circuit may cause calcium-phosphate precipitation in the setting of tumor lysis syndrome. After the procedure, the WBC will rebound without subsequent chemotherapy, and it is not uncommon for two or more procedures to be required to maintain the goal of a WBC count less than50 × 10⁹/L.

Since 1986, the American Society for Apheresis (ASFA) has published evidence-based guidelines for the use of therapeutic apheresis based on stringent review of new literature in the field. Each indication is categorized and supported by the strength of this recommendation based on the quality of evidence available. In 2016, the ASFA published its updated guidelines for the use of leukocytapheresis in the management of patients with hyperleukocytosis. In view of the growing literature at that time suggesting a lack of improvement in early mortality (including the systematic review by Oberoi et al¹²), the recommendation for leukocytapheresis for AML with leukostasis clinical complications was downgraded from a Category I to a Category II, suggesting its use as a second-line treatment option. A major concern was that the previous Category I indication (front-line therapy) might deflect or delay the initiation of induction chemotherapy.¹³ The 2019 ASFA guidelines further qualified the recommendation grade to a weak recommendation with moderate-quality evidence.⁵ Prophylaxis of asymptomatic hyperleukocytosis is a Category III indication (Grade 2C recommendation) for leukocytapheresis in the 2019 ASFA guidelines, and there is no significant evidence to support the use of leukocytapheresis to reduce the renal and metabolic complications associated with tumor lysis syndrome.⁵ The 2015 British Committee for Standards in Haematology guidelines recommend the consideration of leukocytapheresis as part of the initial management of hyperleukocytosis complicated by clinical leukostasis (pulmonary, cerebral, and/or renal leukostasis or priapism) or when the use of chemotherapy is problematic (eg, during pregnancy).¹⁴ They also make a strong recommendation that clinicians must aim for early initiation of suitable chemotherapy and good supportive care irrespective of the use of leukocytapheresis. Given the conflicting data over the years, it is not surprising that a survey in 2018 of Eastern Cooperative Oncology Group members in the United States found that 79.2% of respondents would use leukocytapheresis followed by induction chemotherapy in a patient with AML and hyperleukocytosis with leukostasis.¹⁵ In addition, 32.8% of respondents would use leukocytapheresis without leukostasis. The survey demonstrated widespread variability in the management of hyperleukocytosis and perceptions regarding indications, contraindications, and outcomes with leukocytapheresis.¹⁵

There are no randomized prospective studies on the use of leukocytapheresis in hyperleukocytosis or leukostasis. There will likely never be one, given the lack of accepted diagnostic parameters, validated biomarkers of leukostasis, and uniform criteria for therapeutic endpoints, as well as the logistical difficulties of conducting trials in uncommon, emergent clinical settings. The well-designed and methodologically sound meta-analysis by Bewersdorf et al provides additional strong evidence to discourage the routine adjunctive use of leukocytapheresis in patients with AML with hyperleukocytosis and leukostasis. Certain select populations, such as older individuals or others who are not candidates for induction chemotherapy, could be considered for leukocytapheresis as a bridge to control leukostasis complications while novel or alternative treatments are employed. Under all circumstances, leukocytapheresis is labor and cost intensive, logistically demanding, not universally available, and associated with adverse effects. The priority for patients with AML and hyperleukocytosis should be initiation of “intention to treat” targeted and/or cytoreductive induction chemotherapy along with aggressive supportive care measures for DIC and tumor lysis syndrome.

CONFLICT OF INTEREST

The authors declare no conflicts of interest.