Establishing Pragmatic Analytical Performance Specifications for Blood Beta-Hydroxybutyrate Testing

Abstract

Background: Currently, no authoritative guidelines exist recommending the analytical performance specification (APS) of blood beta-hydroxybutyrate (BOHB) testing in order to meet the clinical needs of patients. This study has applied existing diabetic ketoacidosis (DKA) BOHB diagnostic thresholds and the recommended rates of fall in BOHB concentrations during DKA treatment to establish pragmatic APSs for BOHB testing.

Methods: Required analytical performance was based on 2 clinical requirements: (a) to reliably distinguish between non-adjacent DKA BOHB diagnostic categories of <0.6, 0.6 to 1.5, 1.6 to 2.9, and ≥3 mmol/L, and (b) to be assured that a measured 0.5 mmol/L reduction in BOHB indicates the true concentration is at least falling (meaning >0 mmol/L decline).

Results: An analytical coefficient of variation (CV) of <21.5% could reliably distinguish all non-adjacent diagnostic categories with >99% certainty, assuming zero bias. In contrast, within-day CVs of 4.9%, 7.0%, and 9.1% at 3 mmol/L BOHB were required to assure truly falling ketone concentrations with 99% (optimal), 95% (desirable), and 90% (minimal) probability, respectively. These CVs are larger at lower BOHB concentrations and smaller at higher concentrations.

Conclusions: Reliable tracking of changes in BOHB during DKA treatment largely drives the requirement for analytical performance. These data can be used to guide minimal, desirable, and optimal performance targets for BOHB meters and laboratory assays.

Fig. 1.

Effect of analytical bias on required assay imprecision at a beta-hydroxybutyrate concentration of 3 mmol/L with 99%, 95%, and 90% probability of not miscataloging the result by more than one category.