Cold-vapour atomic absorption spectrometry underestimates total mercury in blood and urine compared to inductively-coupled plasma mass spectrometry: an important factor for determining mercury reference intervals

Abstract

Aim: In recent years, the application of inductively-coupled plasma mass spectrometry (ICP-MS) has been used increasingly in clinical laboratories for the measurement of various trace elements and heavy metals. However, full evaluation of this technique has not been conducted to ensure the transfer of comparable results from conventional cold-vapour atomic absorption spectrophotometry (CVAAS) for blood and urine total mercury (Hg) analysis.

Methods: A total of 131 blood and 223 urine samples from both patients and normal healthy subjects were collected from a university-based trace element laboratory and a population survey of healthy school adolescents. Correlation study was conducted for total Hg concentration measured by the traditional on-line digestion with flow injection CVAAS and the newly installed ICP-MS. Reference materials were used for method validation and quality control. Standard addition of fixed amounts of inorganic and methyl Hg standards into blood and urine were performed for recovery study. Bias in total Hg measurement was investigated by re-calibrating both instruments using methyl Hg standards.

Results: The intra- and inter-assay coefficients of variation in the ICP-MS were <6% in the range of 14-259 nmol/L for Hg in blood and urine samples assayed. The detection limit was 1.1 nmol/L and linearity was up to 186 nmol/L. The results from analyses of a range of whole blood and urine reference materials agreed well with the certified values. The correlation study showed a significant correlation between ICP-MS and CVAAS with: [ICP-MS] = 7.36 + 1.69*[CVAAS] in blood samples (r = 0.84, p < 0.0001) and [ICP-MS] = 1.90 + 1.14*[CVAAS] in urine samples (r = 0.93, p < 0.0001) for total Hg. Recovery study showed that the % recovery of inorganic Hg in blood for ICP-MS and CVAAS ranged from 83 to 95% and 77 to 84%, respectively, while that of inorganic Hg in urine for ICP-MS and CVAAS ranged from 92 to 126% and 43 to 93%, respectively. For methyl Hg, the % recovery in blood for ICP-MS and CVAAS ranged from 72 to 89% and 37 to 75%, respectively, while that in urine for ICP-MS and CVAAS ranged from 65 to 85% and 29 to 42%, respectively. When both instruments were re-calibrated using methyl Hg standards, the blood and urinary total Hg results in ICP-MS were corrected at 24% and -11% of CVAAS, respectively.

Conclusions: Analysis of total Hg was underestimated at about 69% in blood and 14% in urine using the traditional CVAAS method compared to ICP-MS, plausibly due to incomplete oxidation and reduction of methyl Hg species in CVAAS method. The normal limit of blood total Hg concentration has been targeted at <50 nmol/L based on the traditional CVAAS method, and the in vivo proportion of methyl Hg of individuals mainly depends on the dietary intake of seafood. Therefore, for clinical laboratories preparing to change over to ICP-MS method for total Hg analysis, the local reference interval for blood total Hg should be re-determined using a non-occupationally exposed population. Otherwise, over-diagnosis of Hg intoxication can result. We have found that by using ICP-MS for total Hg analysis, the local reference range in blood was <77 nmol/L while in spot urine was <15 nmol/L or 1.2 nmol/mmol of creatinine.