Go, Lewis T

191 Background: ANC is used to determine the dosing of chemotherapy and eligibility in clinical trials. Duffy-null phenotype is associated with a lower ANC and present in 2/3 of the Black population. Therefore, ANC reference intervals from hospitals serving a larger percentage (%) of Black population are expected to be lower. Ideally, reference intervals should represent the normal values of the population served. The objective of this study is to determine if the ANC reference intervals of hospitals in the US are representative of the population served based on racial diversity. Methods: We obtained hospital ANC reference intervals from the Mayo Clinic EPIC electronic medical records Care Everywhere health information exchange. Reference intervals from 2020-2023 were included. For each hospital, we determined the city where it is located. For each city, the % of Black population was obtained from the 2020 US Census. We studied the correlation (Pearson) between reference ANC values and % of Black population. To determine the expected ANC reference intervals, we obtained ANC data from the 2021-2022 National Health and Nutrition Examination Survey (NHANES) and performed a reference interval simulation study. Individuals with cancer and cirrhosis were excluded. Random samples of 120 individuals were selected, consisting of n samples from Black individuals and 120-n samples from non-Black individuals. This was performed for n from 0 to 120, and repeated 1000 times per n. Results: We collected data from 107 hospitals across 40 states and 81 cities. Fifty-three (49.5%) hospitals were academic. The % of Black populations in the cities served by the hospitals was a mean of 23.4% (range, 0.7%-82.8%). There was no correlation between ANC upper (r = -0.17, p=0.07) and lower (r = 0.01, p=0.88) limits based on the % Black population. This finding was similar in both academic and non-academic hospitals. Assessment of the ANC distribution in the US population based on race by using the NHANES database showed a clear association with race at the 2.5th (p<0.001) and 97.5th (p=0.009) percentiles with an overall estimated mid-95% reference range of 2.0-8.3 for non-Black individuals and 1.3-7.7 for Black individuals. Conclusions: Hospitals serving a higher % of Black population have similar upper and lower ANC reference intervals compared to those serving a lower % of Black population. This contrasts with the findings from NHANES data which showed significant variations in the normal ANC values between Black and non-Black individuals. Therefore, current hospital ANC reference intervals may not represent the diversity of the population served. Potential clinical implications include misdiagnosis of neutropenia, delay or underdosing of chemotherapy, and inappropriate exclusion in clinical trials, especially in areas with a high % of Black population.