Severe Leukocytosis in Solid Tumor Patients

Abstract & Commentary

By Andrew S. Artz, MD, Division of Hematology/Oncology, University of Chicago. Dr. Artz reports no financial relationships relevant to this field of study.

Synopsis: Very limited data exist regarding causes and prognosis for extreme leukocytosis among non-hematologic malignancy patients. The authors identified 758 solid-tumor patients with a white blood-cell count exceeding 40,000/uL. The causes included hematopoietic growth factors (69%), infection (15%), corticosteroids (5%), and newly diagnosed leukemia (1%). Thus, 77 (10%) had no cause and were considered to be a paraneoplastic leukemoid reaction (PLR). Among PLR patients, the leukocytosis was primarily a neutrophilia, most had metastatic disease, and 78% died or were enrolled in hospice within 12 weeks. Extreme leukocytosis in solid tumor patients is often related to hematopoietic growth factors. Patients with PLR generally have advanced disease and a poor prognosis.

Source: Granger J, Kontoyiannis D. Etiology and outcome of extreme leukocytosis in 758 nonhematologic cancer patients: A retrospective, single institution study. Cancer. 2009;115:3919- 3923.

Extreme leukocytosis among patients with nonhematologic malignancies is an infrequent but troublesome issue.1,2 The prevalence, causes, and prognostic relevance remain poorly described. Etiologies may include infection, corticosteroids, growth factors, a concurrent hematologic malignancy, an acute inflammatory reaction, or a paraneoplastic leukemoid reaction (PLR).

PLR appears to occur among many tumor types, and is likely driven by colony-stimulating factors such as G-CSF, GM-CSF, interleukin-6 or other growth factors.2-5 Without specific assays, PLR is a diagnosis of exclusion.

The authors reviewed the electronic medical records of cancer patients with a leukocyte count > 40,000/uL over a three-year period. All patients with hematologic malignancy were excluded. Among 3,770 consecutive patients with extreme leukocytosis, 758 had solid tumors and formed the study cohort. Among these, leukocytosis was related to hematopoietic growth factors in 522 (69%), infection in 112 (15%), steroids or vasopressors in 38 (5%), and leukemia in 9 (1%). The remaining 77 patients (10%) without an obvious cause were categorized as related to PLR.

Among those with infections, half were related to pneumonia and 27% had bacteremia. A broad range of infectious agents were identified. For the 77 patients having extreme leukocytosis related to PLR, the mean WBC was 53,000/uL. Leukocytosis was generally neutrophilic. Almost all (99%) were afebrile, and many had documented white blood cell counts > 20,000/uL at least one month before the white blood-cell count exceeded 40,000/uL indicated a subacute process. Most (78%) had metastatic disease and an unspecified number had large primary tumors. Among PLR patients, 54 patients (76%) died within 12 weeks or were enrolled in hospice and 89% died within one year of the initially documented extreme leukocytosis. The 10% who survived for > one year had leukocytosis that resolved after chemotherapy and/or surgical resection of the tumor.


The occurrence of leukocytosis in the setting of a non-hematologic malignancy (i.e., a "solid tumor") is not uncommon clinically, but precise estimates are lacking. While many cases will be clearly associated with infection, steroids, or growth factors, a significant and troublesome proportion may have no obvious cause. PLR may be defined as leukocytosis without another cause and related to the underlying malignancy. As with other paraneoplastic syndromes, factors secreted by the tumor appear to drive the leukocytosis. G-CSF and GM-CSF secretion are best described, but other factors may be involved.2-5 Since no clinical assay is readily available to test the serum or a tumor block, the diagnosis remains a clinical diagnosis of exclusion.

This retrospective review by Granger and Kontoyiannis represents the largest series to date on severe leukocytosis in non-hematologic malignancy patients. They defined extreme leukocytosis as greater than 40,000/uL. They found 758 solid tumor patients with extreme leukocytosis. Hematopoietic growth factors accounted for 69% of cases of severe leukocytosis, with only 11% a documented infection. Concurrent, newly diagnosed leukemia was uncommon (1%). Seventy-seven (10%) patients without infection, leukemia, or exposure to steroids or growth factors were classified as related to PLR. As expected, most of these patients had metastatic disease. The authors state these related to a variety of tumor types but did not specify the exact range. The majority of PLR patients (76%) died within 12 weeks of finding extreme leukocytosis. Interestingly, a subset of 10% survived for > 1 year, all of whom received effective chemotherapy or radiation therapy.

Unfortunately, the authors do not provide the denominator of the number of solid tumor patients evaluable to estimate the prevalence of extreme leukocytosis. Further, we do not know if the review was restricted to inpatients, outpatients, or both. One would expect quite different findings, at least for causes of extreme leukocytosis between inpatients and outpatients. Finally, the survey of a tertiary cancer center may not reflect the typical population an oncologist sees.

In addition, as a retrospective review, diagnostic classification is limited. Confidently excluding infection as a cause is difficult without prospective evaluation. However, the authors appear to have performed a thorough review to exclude alternative causes for leukocytosis so that diagnosing as PLR appears valid. For example, among those with PLR, only one had a fever and most had less severe leukocytosis more than one month before extreme leukocytosis was detected, providing confidence in excluding underlying infection or other reversible causes. Using an extremely high WBC probably enhanced accuracy for PLR, as clinicians would generally perform an infectious evaluation for severe leukocytosis. The results may represent the "tip of the iceberg," in that less severe leukocytosis is certainly more common. Prospective studies of less severe leukocytosis are clearly warranted to confirm the various causes, to determine the prevalence of PLR, and to determine the prognostic value. More importantly, oncologists could use data to support the extent to which other factors must be excluded before diagnosing PLR among a solid tumor patient with leukocytosis.

In conclusion, extreme leukocytosis (> 40,000/uL) among patients with non-hematologic malignancies is often related to hematopoietic growth factors and less often to infection or steroids. However, in around 10% of the cases, leukocytosis is a paraneoplastic reaction and generally is associated with advanced cancer and a poor prognosis.


1. McKee LC, Jr. Excess leukocytosis (leukemoid reactions) associated with malignant diseases. South Med J. 1985;78:1475-1482.

2. Kojima K, et al. Clinicopathological study of involvement of granulocyte colony stimulating factor and granulocyte-macrophage colony stimulating factor in non-lymphohematopoietic malignant tumors accompanied by leukocytosis. Histol Histopathol. 2002;17: 1005-1016.

3. Cvitkovic E, et al. Leukemoid reaction, bone marrow invasion, fever of unknown origin, and metastatic pattern in the natural history of advanced undifferentiated carcinoma of nasopharyngeal type: a review of 255 consecutive cases. J Clin Oncol. 1993;11:2434-2442.

4. Kasuga I, et al. Tumor-related leukocytosis is linked with poor prognosis in patients with lung carcinoma. Cancer. 2001;92:2399-2405.

5. Sawyers CL, et al. Production of granulocyte-macrophage colony-stimulating factor in two patients with lung cancer, leukocytosis, and eosinophilia. Cancer. 1992;69:1342-1346.