Oncofusion

start / graph / concept-hyperleukocytosis-emergency

Explainerstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q7-clinical

Hyperleukocytosis: why a very high white-cell count is a medical emergency

also called hyperleukocytosis · leukostasis · high white cell count at diagnosis · WBC > 100

In KMT2A-rearranged infant leukaemia the blood can carry so many leukaemia cells that it becomes thick and sludgy, and the cells plug the smallest blood vessels in the lungs and brain. That is why a baby with a white-cell count in the hundreds of thousands is treated as an emergency within hours, not days.

The same thing, for a scientist

Hyperleukocytosis is conventionally defined as WBC >= 100 x 10^9/L and is characteristic of KMT2A-rearranged infant ALL: median presenting WBC was 145 x 10^9/L in the COG AALL0631 intermediate-risk group and 213 x 10^9/L in the high-risk (<90 days) group. The acute hazards are pulmonary and cerebral leukostasis, intracranial haemorrhage, tumour lysis syndrome and disseminated intravascular coagulation. In paediatric series, pulmonary leukostasis was present in 21.3% at admission, and cerebral leukostasis/coagulopathy-related early death within 14 days occurred in 4.2% of ALL and 7.7% of AML patients with hyperleukocytosis, all of whom had WBC >= 400 x 10^9/L. Management is hyperhydration, avoidance of transfusion-driven viscosity increases, tumour-lysis prophylaxis and prompt cytoreductive chemotherapy; leukapheresis is used selectively and is not of proven survival benefit.

Go deeper
The physics of the problem1 deeper

Blood is a suspension. Raise the cell fraction far enough and its viscosity rises non-linearly, and the smallest vessels - pulmonary capillaries, cerebral arterioles - are the first to stop flowing. Leukaemic blasts make this worse than an equivalent number of red cells would, because they are large, stiff, and adhesive: they express adhesion molecules and can bind endothelium rather than sliding past it. Two organs pay the price first:

  • Lung. Pulmonary leukostasis presents as tachypnoea, hypoxia and diffuse infiltrates that look like infection or fluid overload. In one paediatric hyperleukocytosis cohort 13 of 61 patients (21.3%) had pulmonary leukostasis at admission, and all recovered with chemotherapy and nasal oxygen 2.
  • Brain. Cerebral leukostasis presents as irritability, altered consciousness, seizures or focal signs, and it can tip into intracranial haemorrhage. In the same cohort, early death within 14 days from cerebral leukostasis or coagulopathy was 4.2% in ALL and 7.7% in AML, and every one of those patients had a WBC of at least 400 x 10^9/L 2.
Why an infant is a harder version of this problem

The counts in KMT2A-rearranged infant leukaemia are among the highest in paediatric oncology: median 145 x 10^9/L in the COG AALL0631 intermediate-risk group (range 2-4335) and 213 x 10^9/L in the high-risk group under 90 days (range 12-1800) 1. A count of 4335 x 10^9/L is more than four hundred times the upper limit of normal. At the same time the patient weighs a few kilograms, so absolute fluid volumes for hyperhydration, apheresis circuit volumes and drug doses are all at the edge of what the equipment was designed for, and an extracorporeal circuit may hold a substantial fraction of the baby's total blood volume.

What the team actually does in the first hours1 deeper
  1. Hyperhydration at roughly twice maintenance, without potassium, to protect the kidneys from the urate and phosphate load of dying cells.
  2. Tumour-lysis prophylaxis: allopurinol, or rasburicase if urate is already high or renal function impaired. In a series managed without routine rasburicase or apheresis, laboratory tumour lysis still developed in 54.2% of ALL and 76.9% of AML patients but no patient required dialysis 2.
  3. Restraint with red-cell transfusion. Correcting anaemia raises haematocrit and therefore viscosity; transfusion is deferred or given slowly unless the child is compromised. Platelets, by contrast, are given freely because they do not materially change viscosity and because intracranial haemorrhage is the main cause of early death 2.
  4. Start cytoreductive chemotherapy early. This, not apheresis, is the definitive treatment - in one centre time from presentation to chemotherapy was a mean of 34-35 hours whether or not apheresis was used 3.
  5. Leukapheresis, selectively. Used in 8 of 62 (13%) hyperleukocytosis patients at one centre, with minimal complications; those selected had much higher counts (mean 498 vs 237 x 10^9/L) and were far more likely to have neurological (63% vs 17%) or pulmonary (75% vs 17%) leukostasis symptoms 3. There is no consensus that it improves survival, and the authors of the leukapheresis-free series argue it can be reserved for patients with leukostasis in whom chemotherapy cannot be started promptly 23.
The honest state of the evidence

Both of the cohorts cited here are single-centre retrospective series of a few dozen patients, in a setting where the sickest patients are selected for the intervention. That is precisely the design that cannot separate the effect of apheresis from the effect of being the kind of patient who gets it. No randomised trial of leukapheresis in children exists, and its absence is itself a finding: the population is too small and too acutely unwell to randomise easily.

Why this matters beyond the first week

Presenting WBC is not only an acute hazard, it is a durable prognostic variable. In Interfant-06, WBC count was an independent prognostic factor for event-free survival after KMT2A rearrangement and age, and WBC >= 300 x 10^9/L in an infant older than 6 months is one of the two criteria that define the high-risk group, whose 6-year EFS was 20.9% (SE 3.4) 4. In the CCCG-ALL-2015 infant cohort 11.2% of infants met that WBC >= 300 threshold 5. So the same number that drives the emergency management also determines which arm of the protocol the baby enters and whether transplant is planned.

Facts and figures
definitionWBC >= 100 x 10^9/L (some series use >= 50 for AML)
max reported wbc x10e9 per L4335
prognostic thresholdWBC >= 300 x 10^9/L defines Interfant-06 high risk in infants > 6 months
median presenting wbc x10e9 per L
cccg_all_2015_infants62.7cccg_all_2015_infants: 62.7cog_aall0631_ir145.0cog_aall0631_ir: 145cog_aall0631_hr213.0cog_aall0631_hr: 213
complications
early death <14 days (cer…4.200early death <14 days (cerebral leukostasis/coagulopathy): 4.2early death <14 days (cer…7.700early death <14 days (cerebral leukostasis/coagulopathy): 7.7clinical tumour lysis syn…14.6clinical tumour lysis syndrome: 14.58pulmonary leukostasis21.3pulmonary leukostasis: 21.3laboratory tumour lysis s…54.2laboratory tumour lysis syndrome: 54.16laboratory tumour lysis s…76.9laboratory tumour lysis syndrome: 76.92
median presenting wbc x10e9 per L — the numbers behind the chart
{
  "cog_aall0631_ir": 145,
  "cog_aall0631_hr": 213,
  "cccg_all_2015_infants": 62.7
}
PCT wbc over 300
{
  "cccg_all_2015_infants": 11.2
}
complications — the numbers behind the chart
[
  {
    "name": "pulmonary leukostasis",
    "frequency_pct": 21.3,
    "cohort": "61 children with hyperleukocytosis"
  },
  {
    "name": "laboratory tumour lysis syndrome",
    "frequency_pct": 54.16,
    "cohort": "ALL with hyperleukocytosis"
  },
  {
    "name": "laboratory tumour lysis syndrome",
    "frequency_pct": 76.92,
    "cohort": "AML with hyperleukocytosis"
  },
  {
    "name": "clinical tumour lysis syndrome",
    "frequency_pct": 14.58,
    "cohort": "ALL with hyperleukocytosis"
  },
  {
    "name": "early death <14 days (cerebral leukostasis/coagulopathy)",
    "frequency_pct": 4.2,
    "cohort": "ALL with hyperleukocytosis",
    "note": "all had WBC >= 400 x 10^9/L"
  },
  {
    "name": "early death <14 days (cerebral leukostasis/coagulopathy)",
    "frequency_pct": 7.7,
    "cohort": "AML with hyperleukocytosis"
  }
]
management
[
  "hyperhydration without potassium",
  "allopurinol or rasburicase",
  "defer/slow red-cell transfusion",
  "liberal platelet transfusion",
  "early cytoreductive chemotherapy",
  "leukapheresis in selected patients with leukostasis"
]
How this could bite us back
  • Rapid cytoreduction in a patient with a huge tumour burden is exactly the setting for tumour lysis syndrome and, for menin inhibitors specifically, for differentiation syndrome - so an effective new agent creates its own acute hazard at presentation.
  • If a targeted agent is used for initial cytoreduction before the diagnostic samples are complete, the leukaemia may be altered or cleared before the fusion partner and the genomic breakpoint (the MRD tracking target) have been characterised.
  • Anything that lowers the count without killing the self-renewing compartment will make the patient look better while leaving the disease intact, and presenting WBC is one of the variables used to assign risk group - so a cosmetic response could reassign a high-risk baby into a lower-risk arm.
Still open
  1. Is there any threshold or symptom profile at which leukapheresis demonstrably reduces early death in infants, or is the whole effect attributable to getting chemotherapy started sooner?
  2. Does the very high presenting WBC of KMT2A-rearranged infant leukaemia reflect a higher proliferative rate, a failure of blasts to be retained in the marrow niche, or an altered adhesion phenotype - and would a menin inhibitor reduce it fast enough to be useful as initial cytoreduction?
  3. Could an oral targeted agent given in the first 24-48 hours provide safer cytoreduction than a steroid prophase in a baby with a count above 400 x 10^9/L, where intracranial haemorrhage is the dominant cause of early death?
Every source behind this entry (5)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In COG AALL0631, median presenting WBC among KMT2A-rearranged infants was 145 x 10^9/L (range 2-4335) in the intermediate-risk group (age 90-365 days, n = 111) and 213 x 10^9/L (range 12-1800) in the high-risk group (age < 90 days, n = 35).

    PMID 33623141DOI
    What this evidence cannot show

    A KMT2A-rearranged-only trial cohort, so these medians are higher than those of all-comer infant ALL cohorts (the CCCG-ALL-2015 infant median was 62.7 x 10^9/L including KMT2A-germline cases). Ranges are extreme and the distribution is heavily right-skewed, so the median understates how often a count is dangerous.

    Evidence that argues against this (1)
    • A median with a range from 2 to 4,335 is a distribution so skewed that the median is a poor summary and the difference between 145 and 213 across risk groups is within what such a distribution produces by chance. The younger group also has 35 patients. Statistics in Medicine 2006
  2. E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In a paediatric cohort of 61 children with hyperleukocytosis (48 ALL, 13 AML) managed without leukapheresis and without routine rasburicase, 13 (21.3%) had pulmonary leukostasis at admission and all recovered with chemotherapy and nasal oxygen; laboratory tumour lysis developed in 54.16% of ALL and 76.92% of AML patients and clinical tumour lysis in 14.58% and 15.38%, with no patient requiring dialysis; early death within 14 days from cerebral leukostasis or coagulopathy was 4.2% in ALL and 7.7% in AML, and all such patients had a WBC >= 400 x 10^9/L; intracranial haemorrhage rather than renal failure was the main cause of early death.

    Management of Hyperleukocytosis in Childhood Acute Leukemia Without Leukapheresis and Rasburicase Prophylaxis · Aylan Gelen S, Sarper N, Zengin E, et al. · Journal of Pediatric Hematology/Oncology 2022
    PMID 34054042DOI
    What this evidence cannot show

    Retrospective, single centre, 61 patients, in a setting where leukapheresis was unavailable rather than deliberately withheld - so this is an observational demonstration that management without apheresis is feasible, not evidence that apheresis is useless. The cohort is children in general, not infants, and infants are not reported separately.

    Evidence that argues against this (1)
    • A centre that does use leukapheresis reports it in 13% of hyperleukocytosis patients with minimal complications, and reserves it for those with neurological or pulmonary leukostasis, which is a defensible alternative position. source 2022
  3. E3weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper

    At a single institution over 10 years, leukapheresis was used in 8 of 62 (13%) paediatric patients with hyperleukocytosis with minimal complications; mean leucocyte count was 498 x 10^9/L in those apheresed versus 237 x 10^9/L in those not; patients with neurological (63% vs 17%) or pulmonary (75% vs 17%) leukostasis symptoms were more likely to be apheresed; time from presentation to chemotherapy was unchanged (mean 35 vs 34 hours); and the one death in the apheresis group was from neurological sequelae of hyperleukocytosis rather than from the procedure.

    PMID 35455547DOI
    What this evidence cannot show

    Retrospective case-control, 8 apheresed patients. Apheresis was given to the sickest patients with the highest counts, so no outcome comparison between the groups can be causal. The paper supports only the narrow claim that apheresis was safe and did not delay chemotherapy at that centre. No infant-specific data.

    Evidence that argues against this (2)
    • Eight patients at one institution over ten years, and the comparison is confounded by indication in the strongest way: those apheresed had roughly twice the leucocyte count and far more leukostasis symptoms, which is why they were selected. Systematic review does not support the practice — leukapheresis and low-dose chemotherapy do not reduce early mortality in hyperleukocytic AML. Leukemia Research 2014
    • A later systematic review and meta-analysis reaches the same place: leukapheresis for hyperleukocytosis in AML has no demonstrated survival benefit, and the evidence base is retrospective and confounded throughout. The paediatric evidence is thinner still. Transfusion 2020
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    In Interfant-06, WBC count was an independent prognostic factor for event-free survival after KMT2A rearrangement and age; the high-risk group was defined as KMT2A-rearranged infants older than 6 months with WBC >= 300 x 10^9/L or a poor prednisone response, and had 6-year EFS of 20.9% (SE 3.4).

    PMID 31283407DOI
    What this evidence cannot show

    The WBC >= 300 threshold is a protocol convention; the underlying relationship with outcome is continuous, so a baby at 280 x 10^9/L is not biologically different from one at 310. The high-risk EFS also reflects that only 46% of these infants actually received their intended transplant.

    Evidence that argues against this (1)
    • Independent prognostic significance after KMT2A status and age is a statistical statement about a particular model on a particular cohort, and the 300 x 10^9/L threshold inside the risk definition was chosen rather than estimated. A 6-year EFS of 20.9% in the resulting stratum measures how the strata were drawn as much as how the patients fared. Statistics in Medicine 2006
  5. E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In the CCCG-ALL-2015 infant cohort (n = 134), median WBC was 62.7 x 10^9/L (range 1.3-777) and 11.2% of infants had a WBC of at least 300 x 10^9/L.

    PMID 38634141DOI
    What this evidence cannot show

    Includes KMT2A-germline infants, which lowers the median and the proportion above 300 relative to a KMT2A-rearranged-only cohort.

    Evidence that argues against this (1)
    • CCCG-ALL-2015 is a different population and a different era: 134 infants of whom 53% are KMT2A-rearranged, median age 254 days. Its median WBC of 62.7 against 145 in AALL0631 largely reflects that nearly half the cohort does not have the rearrangement, so the two figures are not comparable as written. Leukemia 2021
Connected to