The differential diagnosis of basophilia in patients undergoing BCR-ABL testing

To the Editor:

Basophilia is an uncommon finding of the peripheral blood detected on complete blood count analysis. Among peripheral blood leukocytes, the basophil is the least abundant, comprising less than 1.0% of peripheral white blood cells.¹ Basophilia is defined by an absolute basophil count greater than the upper reference limit, typically around 0.1 × 10⁹/L. Little is known regarding the prevalence of basophilia in the general population, though a prior study noted that peripheral basophilia was present in 3.0% of patients undergoing bone marrow aspiration.²

Basophilia is associated with myeloid neoplasms, most notably chronic myelogenous leukemia (CML).³ Patients with persistent basophilia detected on peripheral blood often receive testing for CML by assessing for the presence of the BCR-ABL gene rearrangement by either polymerase chain reaction (PCR) or fluorescent in situ hybridization (FISH). Though the association of basophilia and CML is well recognized, little is known regarding other conditions that may be related to this peripheral blood abnormality. Our aims were to investigate the differential diagnosis of basophilia in patients with leukocytosis who underwent diagnostic evaluation for CML, and evaluate the relationship between the degree of basophilia and subsequent diagnosis.

We reviewed the medical records of patients who underwent BCR-ABL testing at the Mayo Clinic from 1 January 2014 to 31 December 2018. Complete blood count analysis at time of BCR-ABL testing was assessed and patients who had both leukocytosis and absolute basophilia were included. Individuals who already carried a diagnosis of CML were excluded. Leukocytosis was defined as a total white blood cell count ≥9.7 × 10⁹/L. Basophilia was defined as an absolute basophil count of ≥0.09 × 10⁹/L per institutional lab parameters.

Three-hundred eighty-two patients with leukocytosis and basophilia and without prior CML diagnosis underwent BCR-ABL testing. Two-hundred one patients were male (52.6%) and the median age was 59 years (range: 19 to 89 years). The median total white blood cell count was 16.1 × 10⁹/L (range: 9.7 to 746.5 × 10⁹/L). The median absolute basophil count was 0.20 × 10⁹/L (range: 0.09 to 54.5 × 10⁹/L).

Study results are summarized in Table 1. Myeloid neoplasms were diagnosed in 191 patients (50%): CML (104), essential thrombocythemia (23), polycythemia vera (21), primary myelofibrosis (18), unclassifiable myeloproliferative neoplasms (10), chronic myelomonocytic leukemia (8), acute myelogenous leukemia (4), chronic neutrophilic leukemia (2), and myelodysplastic syndrome (1). In 70 patients (18.3%), a hematologist identified a comorbid condition that coincided with the patientʼs basophilic leukocytosis: malignancy (20), tobacco use (15), corticosteroid use (7), infection (6), autoimmune disease (5), prior splenectomy (5), monoclonal gammopathy of undetermined significance (3), and other inflammatory states (9). No comorbid condition was identified in 121 patients (31.7%). In patients in whom basophilia was deemed to be reactive to another comorbid condition, resolution of basophilia occurred in 75.4% of patients who received follow up blood count analysis.

TABLE 1.

Patient information and diagnoses noted to coincide with basophilia

Patient demographic
Median age (y)	59 (19–89)
Male sex	201 (52.6%)
Median absolute basophil count	0.20 × 109/L (0.09–54.5 × 109/L)
Condition	Number of patients
Myeloid neoplasm	191
Chronic myeloid leukemia	104
Essential thrombocythemia	23
Polycythemia vera	21
Primary myelofibrosis	18
Unclassifiable myeloproliferative neoplasm	10
Chronic myelomonocytic leukemia	8
Acute myeloid leukemia	4
Chronic neutrophilic leukemia	2
Myelodysplastic syndrome	1
Reactive basophilia	191
Malignancy	20
Acute lymphocytic leukemia	7
Solid tumor	5
Chronic lymphocytic lymphoma	3
Nodular sclerosing hodgkin lymphoma	1
Multiple myeloma	1
Peripheral T-cell lymphoma	1
Marginal zone lymphoma	1
Lymphoplasmacytic lymphoma	1
Tobacco use	15
Corticosteroid use	7
Infection	6
Autoimmune disease	5
History of prior splenectomy	5
Monoclonal gammopathy of undetermined significance	3
Hyper-eosinophilic syndrome	2
Systemic mastocytosis	2
Gastrointestinal hemorrhage	2
Mast cell activation syndrome	1
Hepatic steatohepatitis	1
Post-surgical inflammation	1
Idiopathic	121

We then compared the patients with and without myeloid neoplasia in order to assess whether the degree of basophilia was related to the final diagnosis. Patients with myeloid neoplasms had a higher absolute basophil count compared with the remainder of the cohort (Mean: 4.1 × 10⁹/L vs 0.14 × 10⁹/L; P < .01). All patients with an absolute basophil count of ≥0.48 × 10⁹/L had a myeloid neoplasm. The sensitivity and specificity of an absolute basophil count >0.40 × 10⁹/L for myeloid neoplasm was 67.5% (95% confidence interval [CI], 60.4%−74.1%) and 99.0% (95% CI, 96.3%−99.9%) respectively.

Our observations support the notion that in the absence of CML, patients with basophilia may benefit from further evaluation into other myeloid cancers. Prior studies have noted a relationship between basophilia and other myeloproliferative neoplasms, including primary myelofibrosis, essential thrombocythemia, polycythemia vera, and myelodysplastic syndrome.^4,5 However, these studies were limited by the small sample size of patients with non-CML myeloid disease, and only reported an association of basophilia with a narrow scope of myeloid neoplasms. Our study notes a relationship between basophilia and a broad spectrum of other myeloid malignancies, including chronic myelomonocytic leukemia, chronic neutrophilic leukemia, and acute myelogenous leukemia.

Among patients in whom basophilia was considered potentially reactive to a comorbid condition, there was a large variety of noted diagnoses. Less information is known regarding reactive basophilia though prior small studies have identified a higher absolute basophil count in patients with atopy, iron deficiency, and diabetes.^3,6 Interestingly, lymphoid malignancies and tobacco use were the most common non-myeloid disease states coinciding with basophilia. Given that over 75% of patients with reactive basophilia in this study had resolution of basophilia with continued monitoring, it may be reasonable to obtain repeat blood counts in patients in whom a potential comorbid condition is identified.

Patients with myeloid neoplasm had higher absolute basophil counts when compared to the remainder of the cohort. The likelihood of a myeloid malignancy increased as the degree of basophilia increased. Though prior expert opinions have recommended a high index of suspicion for chronic myeloid leukemia in patients with an absolute basophil count >1.0 × 10⁹/L, our study suggests a lower threshold for concern.³ All patients with an absolute basophil count of ≥0.48 × 10⁹/L (n = 122) were noted to have a myeloid neoplasm and the specificity of an absolute basophil count >0.40 × 10⁹/L for myeloid malignancy was 99.0%. Therefore, we recommend evaluation for CML and other myeloid neoplasms in patients with leukocytosis and an absolute basophil count >0.40 × 10⁹/L.

In conclusion, among those undergoing BCR-ABL testing for the evaluation of leukocytosis, only one in four patients with basophilia were found to have CML. The rest were associated with several diseases, a substantial proportion of which were other myeloid neoplasms. All patients with an absolute basophil count ≥0.48 × 10⁹/L had a myeloid malignancy. Though prior studies have suggested a higher absolute basophil count threshold of concern for CML, this study recommends testing for myeloid neoplasms in patients with an absolute basophil count of >0.40 × 10⁹/L.