Multinational Assessment of Absolute Neutrophil Counts and White Blood Cell Counts Among Healthy Individuals with the Duffy Null Phenotype or Genotype

Introduction: The global variation in observed absolute neutrophil count (ANC) among healthy people is largely driven by the rs2814778 CC variant which results in the Duffy null phenotype. This variant results in lower circulating neutrophils, but no increased risk for infection. The Duffy null phenotype is seen in 80% to 100% of people in West Africa and <1% of people of European genetic ancestry. A recent study from the United States proposed a Duffy null specific ANC reference interval of 1,210-5,390/uL. However, it is unclear if this is generalizable to people with this variant from different geographic regions or other self-identified racial/ethnic groups. This multinational study aimed to compare Duffy null specific reference ranges from healthy participants across diverse geographical regions.

Methods: Three prospectively collected datasets of individuals with the Duffy null phenotype from blood donors in Saudi Arabia (n=150) and Namibia (n=122) and healthy primary care patients in Boston (n=120) were created. Participants with the rs2814778 CC variant by genotype from the UK Biobank (n=5917) and the Mass General Brigham (MGB) Biobank (n=1563) were also assessed. Biobank participants with ICD-10 diagnoses that could impact ANC such as cancer, autoimmune disease, or immunodeficiencies were excluded. ANC, white blood cell count (WBC), age, sex, and self-reported race/ethnicity were recorded. Categorical variables were summarized by counts and percentages. Comparisons across groups were estimated using Wilcoxon rank sum test. CLSI guidelines and EP Evaluator version 12.0.0.11 were used to establish a 95% reference range by a nonparametric percentile method with the lower limit of normal (LLN) at 2.5% and upper limit of normal (ULN) at 97.5%.

Results: Calculated Duffy null ANC reference ranges are 820/uL (95% CI: 550-1290) – 6,370/uL (95% CI: 5320-7230) for Namibia, 1,090/uL (95% CI: 930-1440) – 5,100/uL (95% CI: 4670-6180) for Saudi Arabia, and 1,210/uL (95% CI: 1080-1390) – 5,390/uL (95% CI: 4640-5950) for Boston. 21.7% (n=26) of the Boston participants, 27.9% (n=34) of the Namibian participants, and 50.7% (n=76) of the Saudi Arabian participants had ANC below local institutional reference ranges. There is no significant difference in mean ANC between Boston and Namibia (p = 0.512) nor Namibia and Saudi Arabia (p = 0.512). There is a significant difference in mean ANC between Boston and Saudi Arabia (p = 0.031). The ANC reference range for those with the rs2814778 CC variant in the UK Biobank is 1,300/uL - 5,160/uL and 1,190/uL - 10,910/uL in the MGB Biobank.

Duffy null WBC reference range is 2.51 (95% CI: 2.27-3.21) – 9.85 (95% CI: 8.77-11.36) × 10⁹/L for Namibia, 3.71 (95% CI: 2.60-4.02) – 9.95 (95% CI: 9.13-10.98) × 10⁹/L for Saudi Arabia, and 3.00 (95% CI: 2.69-3.57) -9.66 (95% CI: 8.04-11.89) × 10⁹/L for Boston. 10% (n=12) of participants in the Boston cohort, 18% (n=22) of the participants in the Namibia cohort, and 12.7% (n=19) of the Saudi Arabia participants had WBC less than the current institutional reference ranges. There is no significant difference in WBC between Boston and Namibia (p=0.561). There is a difference in WBC between Namibia and Saudi Arabia (p=0.005) and between Boston and Saudi Arabia (p=0.011). For both the ANC and WBC data, there is an overlap of 95% confidence intervals surrounding the LLN and ULN indicating that there is no significant difference in ULN or LLN between any of the prospectively collected cohorts.

Discussion: This multinational effort from prospectively collected data as well as data from biobanks highlights the expected ANC and WBC among healthy people with the rs2814778 CC variant across different continents and racial or ethnic groupings. To our knowledge, this is the first report of expected WBC reference ranges for those with this variant. While we found that there are significant differences in mean ANC and WBC between some prospectively collected cohorts, there are no significant differences in ULN or LLN between cohorts which are the clinically useful values. This emphasizes that it is Duffy status - not country of origin or racial/ethnic groupings - that drives differences in ANC and WBC. Additionally, a significant proportion of Duffy null participants had ANC or WBC below the current institutional LLN. Developing reference intervals by Duffy status is essential to interpret ANC and WBC in Duffy null individuals and accurately determine health.