A robust analysis of alternate methods for estimating blood volume

Abstract

Aim

This study evaluates the accuracy of a newly derived blood volume estimation formula based on the Boer equation for lean body mass, comparing its performance against the Nadler, Allen and Lemmens-Bernstein-Brodsky formulas.

Methods

Blood volume estimation was evaluated using two datasets: the Retzlaff dataset, based on 78 healthy individuals, and the Allen dataset, derived from 81 subjects, two of European descent, the remainder Chinese ‘medical, nursing and pedagogic students, technicians, clerks and family members’ and one young Chinese physician. The formulas were compared using robust statistical methods, including the Wilcoxon Signed-Rank Test, permutation tests, Bland-Altman analysis, and Proportion Within Range.

Results

Across all methods, the formula derived from the Boer equation showed the narrowest limits of agreement and smallest variability in most metrics, highlighting its potential as the most accurate and clinically useful tool for blood volume estimation. The Nadler formula also performed well but with slightly larger errors and bias.

Conclusion

This study highlights the limitations of the Allen formula and demonstrates the superior performance of the Boer formula, which is derived from lean body mass. While the Allen formula performed well on its original dataset, it showed higher variability and less accuracy on more modern data. Both the Nadler and Boer formulas exhibited greater precision, with the Boer formula showing slightly lower variability. The study emphasizes the importance of using independent data sets for validation and addresses a critical gap in blood volume assessment by using robust techniques for analysis.