RESUMEN
BACKGROUND: Discrepancies in ionised calcium concentrations between results from several point-of-care devices derived from intensive care unit (ICU) patients were postulated to be caused by perchlorate, a thyroid blocking agent. The deviations were serious concerning the diagnosis of hyper- or hypocalcaemia and administration of calcium infusions. METHODS: The problem was studied from three perspectives. First: quantification of ionised calcium in heparinised blood samples spiked with sodium perchlorate from healthy volunteers measured using five blood gas analysers (BGAs from IL, Radiometer, Roche and Siemens). Second: verification of clinical concentrations of perchlorate in blood samples after routine use for blood gas analysis from ICU patients. Third: retrospective analysis of data stored from patients during their stay in general ICU of the Departments of Anesthesiology and General Surgery by a patient data management system. RESULTS: Category 1: three of the point-of-care testing devices measure clinically relevant falsely low ionised calcium concentrations when exposed to concentrations perchlorate >0.1 mmol/L. Two were not able to identify hypercalcaemia of up to 4 mmol/L ionised calcium when specific perchlorate concentrations are exceeded. Category 2: measured clinical concentrations of perchlorate ranged between the lower limit of quantification [LOQ=0.03 mmol/L and 2.75 mmol/L (median=0.29 mmol/L; mean=0.585 mmol/L)]. A concentration above 0.1 mmol/L perchlorate was found in 74.3% of all samples that tested positive. Category 3: between February 2006 and July 2009, 42 patients per year (2.2%) received sodium perchlorate with a median length of treatment of 25 days. CONCLUSIONS: Perchlorate causes clinically relevant lowering of measured ionised calcium in BGAs from two providers at concentrations obtained in samples from ICU patients (affecting about 1.1% of all ICU patients).