Establishing multianalyte thresholds for automated detection of EDTA contamination.

Abstract

Introduction: Bias in analytes due to EDTA contamination is easy to miss and represents a common and clinically relevant preanalytical error. We designed a study to determine the in vitro effects of rising rates of EDTA contamination in blood samples and to evaluate how these findings could inform the development of multianalyte thresholds for detecting such errors automatically.

Methods: We incrementally spiked blood samples with increasing volumes of EDTA‑contaminated specimen and measured the resulting changes in alkaline phosphatase, calcium, magnesium, potassium, and iron levels. Using these observations, we developed a series of multianalyte thresholds and evaluated their performance against 1 year of retrospective clinical data (n = 2407).

Results: Our experiments demonstrated that increasing rates of EDTA contamination produced changes in several analytes. Accordingly, we applied multianalyte thresholds using potassium and calcium to retrospective samples. Across all samples, smaller magnitudes of analyte deviation resulted in higher detection rates of subtle EDTA contamination but poorer specificity. We selected a rule of a potassium level of 6.5 mmol/L or higher and a calcium level below 6.5 mg/dL (1.63 mmol/L), which identified 14 samples with a contamination rate of 42.8%.

Discussion: Multianalyte thresholds show promise as an effective approach for automatically identifying clinically significant EDTA contamination in clinical specimens.