(P13-025-24) Evaluation of the Relationship Between Hemoglobin and Hematocrit in Young Children, Women of Reproductive Age, and Pregnant Women in the United States (NHANES 1999-2020)

Objectives: Both hematocrit and hemoglobin are used to identify anemia. Hematocrit (%) is often defined as three times the hemoglobin concentration (g/dL). We evaluated the consistency of this relationship in healthy young children, non-pregnant women of reproductive age (WRA), and pregnant women in the United States.

Methods: Using pooled data from the nationally representative National Health and Nutrition Examination Survey (NHANES) 1999-2020, we performed unadjusted linear regression of hematocrit as a function of hemoglobin in young children (1- < 5 years), non-pregnant WRA (15-49 years), and pregnant women. Regression models were weighted and accounted for the complex survey design. Analyses were limited to healthy participants with white blood cell count ≤10,000 cells/L, C-reactive protein ≤5 mg/L, and normal ferritin levels (>15 µg/L and ≤150 µg/L). Only white blood cell count was collected in all NHANES cycles and was the primary exclusion criteria. C-reactive protein and ferritin levels were used for exclusion when available. Hemoglobin data were not adjusted for elevation. We used unrestricted random sampling to generate population hematocrit-hemoglobin regression coefficients in 500 replicate samples, and tested whether the median slope was equal to 3, using one-sample median signed rank tests.

Results: In healthy young children (n = 4,423), the estimated hematocrit was 2.61 (95% confidence interval (CI) 2.55, 2.67) times hemoglobin.  In healthy WRA (n = 12,119), hematocrit was 2.64 (95% CI 2.60, 2.67) times hemoglobin. In healthy pregnant women (n = 686), hematocrit was 2.76 (95% CI 2.66, 2.86) times hemoglobin. All estimates were significantly different from the standard conversion factor of 3 (p < 0.0001).

Conclusions: In all healthy populations, the unadjusted relationship between hemoglobin and hematocrit was significantly different from 3. Therefore, hematocrit cannot be accurately estimated from hemoglobin using the standard threefold conversion factor, which could lead to differences in anemia identification at the individual level and anemia prevalence at the population level.

Funding Sources: No funding was used for this analysis.