<i>Et tu, brute force?</i> An approach to refine blood sample collection protocols

Reducing the number of samples collected on individual animals during an experiment has various advantages; however, it is critical that any refinements to a sampling protocol retain a similar degree of accuracy. The objective of this study was to develop a refined blood sampling protocol capable of reproducing the original findings of a pilot study using only a subset of the collected samples. Data from a previous study were available consisting of plasma concentrations of Cr and Co collected at 15 timepoints over a 64 h period that were used to calculate area under the curve (AUC). Prior to subsequent analysis, it was determined that an adequate reduced sampling protocol must estimate the original AUC within 10%. A SAS program was prepared that generated all possible combinations of sample timepoints. Optimal sets of timepoints were identified by minimal root mean square prediction error. Output from this procedure can be evaluated for bioequivalence using the two one-sided tests approach. This method highlights the ability to further refine biological sampling protocols, which can be readily applied to other experimental settings.