Rethinking neutropenia thresholds: Ethnicity-specific distributions and mortality associations in the United States adult population using national health and nutrition examination survey 2017–2018

Abstract Introduction Absolute neutrophil count (ANC) is widely used to define neutropenia, with an ANC <1500/μL serving as a universal threshold for clinical risk, chemotherapy eligibility, and trial enrollment. However, emerging evidence suggests this cutoff may not accurately reflect infection or mortality risk across diverse populations. In particular, variation by race/ethnicity has raised questions about the appropriateness of a one-size-fits-all definition. Whether a race/ethnicity-specific definition of neutropenia may be more appropriate remains uncertain. To address this, we analyzed the association between ANC and all-cause mortality in a nationally representative U.S. cohort, stratified by ethnicity, cancer history, and cardiometabolic comorbidities. Methods We used data from the 2017–2018 National Health and Nutrition Examination Survey (NHANES), linked to mortality outcomes through 2019. Adults aged ≥20 years with available ANC and mortality data were included. ANC was derived from complete blood counts, and ethnicity was self-reported. Mortality was assessed using the NCHS Linked Mortality File. We examined ANC distributions and neutropenia prevalence by ethnicity, and investigated the association between ANC with different cutoffs (<500, <1000, <1500) and all-cause mortality using logistic regression. Due to such relationship potentially confounded by comorbidities, we also conducted subgroup analysis stratified by cancer history, diabetes, smoking, heart attack (MI), congestive heart failure (CHF), and stroke. Results Among 5,024 U.S. adults in NHANES 2017–2018, the mean ANC varied significantly by ethnicity. Non-Hispanic Black (NHB) participants had the lowest mean ANC (3.67 ×10³/μL), followed by Other Race/Multi-Racial (OR/MR) (4.16), Mexican American (4.46), and Non-Hispanic White (NHW) (4.55). The prevalence of ANC <1500 was highest in NHB individuals (3.7%). Mortality differed by both ANC category and race. Among NHB participants, the absolute risk of death was 25.0% for those with ANC 500–999, 5.1% for ANC 1000–1499, and 2.1% for ANC ≥1500. In contrast, NHW participants showed 4.1% mortality rate with ANC ≥1500, but no deaths in lower ANC categories. OR/MR participants experienced 20.0% mortality with ANC 1000–1499 and <1% at higher ANC. In race-stratified logistic regression models, ANC <1500 was significantly associated with increased mortality among NHB (OR 4.14, p = 0.033) and OR/MR (OR 15.12, p = 0.028) participants. Specifically, ANC <1000 was independently associated with mortality among NHB adults (OR 16.89, p = 0.024). We then explored whether the association between ANC <1000 and mortality varied by comorbidity burden. In the overall population, ANC <1000 was significantly associated with increased mortality in individuals without cancer, diabetes, CHF, stroke, MI. For example, among participants without diabetes, ANC <1000 was associated with significantly increased mortality (OR 26.64, p = 0.005). Among NHB participants, similar patterns were observed. Notably, the only comorbidity-positive subgroup with a statistically significant link between ANC <1000 and mortality was NHB smokers (n = 497), in whom the risk was markedly elevated. In a multivariable adjusted mortality prediction model, ANC <1000 was associated with 3.35-fold higher odds of mortality (95% CI: 0.28–40.01), and ANC 1000–1499 with 1.90-fold higher odds (95% CI: 0.51–7.09), both relative to ANC ≥1500, though neither reached significance. Independent predictors of mortality included age (OR 1.05 per year), RDW (OR 1.25), WBC (OR 1.10), and prior MI (OR 2.38), while platelet count was protective (OR 0.99 per 1000/μL). Hemoglobin and BMI were not significant in adjusted models. Conclusions In this nationally representative cohort, the widely used ANC <1500 threshold for neutropenia underestimated risk in several populations and may not reflect mortality risk equivalently across subgroups. Our findings support the prognostic relevance of lower ANC cutoffs, particularly <1000 cells/μL, in predicting mortality especially in NHB individuals. These results emphasize the potential of refining neutropenia definition that incorporates ethnic background and clinical context, rather than applying a universal threshold. Further research is needed to inform personalized ANC reference ranges and eligibility criteria in both clinical care and trial design.