Disparities in the Diagnostic Evaluation of Asymptomatic “Neutropenia” and Identification of Duffy-Null Phenotype

Background

Duffy antigens are expressed on erythrocytes and function as a receptor for Plasmodium vivax. Many individuals with ancestry in malaria endemic areas, including Sub-Saharan Africa and the Arabian Peninsula, have a genetic variation in ACKR1 that confers resistance to malaria, which results in the Duffy-null phenotype (Fy[a-,b-]). This is associated with fewer peripheral neutrophils and subsequently lower absolute neutrophil count (ANC) as compared to most laboratory references ranges which are based on individuals of European descent. This was historically called “benign ethnic neutropenia,” and is now known as Duffy-null associated neutrophil count (DANC) to more accurately reflect this normal variant.

Under recognition of Duffy null status has been associated with unnecessary testing and specialty referrals. We aimed to characterize rates of testing and identification of DANC, expected Duffy-null ANC ranges, and clinical outcomes among referrals to Hematology for asymptomatic “neutropenia” in a single center retrospective cohort.

Methods

We queried the UCSF De-identified Clinical Data Warehouse to identify clinical notes of children and adults referred to hematology for asymptomatic “neutropenia,” defined as an ANC below institutional reference ranges without a history of recurrent or atypical infections or oral ulcers, between 2006 and 2024. Notes were subsequently filtered to include those authored by a hematologist/oncologist and mentioning either benign ethnic neutropenia, “Duffy null”, or (Fy[a-,b-]). Deidentified demographic data including age, gender, and self-reported race/ethnicity, diagnostic testing, hematology recommendations, and final outcomes were recorded. Laboratory values were extracted to identify ANC at referral, nadir, and peak values.

Results

252 individuals met inclusion criteria. The median age was 39.5 years (IQR 20-59) and there was an equal distribution of males and females. Individuals who identify as Black comprised the largest group, with 38.8% of the retrospective cohort. RBC antigen testing was recommended in 94 patients, of whom 66 underwent testing and had available results. 49 individuals (19.4%) had confirmed Duffy-null phenotype and 87 individuals (34.5%) were presumed to have DANC without confirmatory testing. Less common diagnoses included drug-induced neutropenia (5.6%), infection-related neutropenia (3.6%), autoimmune neutropenia (3.2%), and hematologic malignancy (2.8%). Median ANC at time of referral was 940/µL (IQR 770-1230; range 400-1650) among Duffy-null individuals and 1080/µL (IQR 537.25-1777.5, range 350-3240) among Duffy non-null individuals. The median ANC nadir in the Duffy null cohort was 790/µL (IQR 660-910, range 316-1650), compared with 945/µL (IQR 537.25-1301.75, range 280-1810) in the Duffy non-null cohort. At the time of referral, 92% of individuals with confirmed DANC had an ANC < 1500/µL, 49% had an ANC < 1000/µL, and 6% had an ANC < 500/µL. 43 patients underwent a diagnostic bone marrow biopsy, of whom 10 individuals (23.3%) were ultimately found to have DANC. 27.9% of patients who underwent bone marrow biopsy were Black despite comprising only 6.6% of the local population in the San Francisco Bay Area (p = 0.002). Among other diagnostic studies, 118 patients had a peripheral smear, 75 had vitamin levels evaluated, and 21 underwent an autoimmune workup. 21 patients had negative impacts on other aspects of health care: 7 individuals had delayed procedures, 6 had medication holds or dose reductions for autoimmune diseases, 4 were deferred for liver or kidney donation, 3 had a delay in chemotherapy, and 1 was excluded from a clinical trial. Of this group, 47.6% were Black identifying. There was significant variability among recommendations for frequency of CBC monitoring and ANC thresholds for bone marrow biopsy.

Conclusions

Black identifying children and adults with ANC lower than established reference ranges are referred to Hematology, undergo invasive testing, and have negative impacts on care at disproportionately higher rates as compared to other groups. We demonstrate here how lack of inclusive ANC reference ranges can lead to over testing and iatrogenic harm. Our data also highlight the utility of RBC antigen testing and the need for further investigation to develop guidance on when additional diagnostic workup is indicated.

No relevant conflicts of interest to declare.