Monitoring Mn in whole blood and urine: a comparison between electrothermal atomic absorption and inorganic mass spectrometry

Measurements of Mn in blood and urine are used in biomonitoring studies as biomarkers of exposure. We compared methods for determining Mn, including graphite furnace atomic absorption spectrometry (GFAAS), quadrupole-based inductively coupled plasma mass spectrometry (Q-ICP-MS) (standard and dynamic reaction cell (DRC) mode), and sector field (SF-) ICP-MS. We investigated polyatomic interferences in ICP-MS that can affect Mn measurements in blood and urine matrices, especially at concentrations associated with biomonitoring studies. Method detection limits (3SD) for blood (and urine) Mn were 1.5 (2.1) mu g L(-1) for GFAAS, 0.6 (0.5) mu g L(-1) for SF-ICP-MS, 1.0 (0.5) mu g L(-1) for DRC-ICP-MS and 6.4 (0.6) mu g L(-1) for Q-ICP-MS operated in standard mode. Method reproducibility for blood (22 mu g L(-1) Mn) and urine matrices (9 mu g L(-1) Mn) was found to be between 3 and 4% RSD for all methods, except for blood Mn by Q-ICP-MS (11% RSD) and urine Mn by GFAAS (7% RSD). Accuracy was assessed using various reference materials. Careful optimization of the DRC mode is required with blood because, in addition to polyatomic interferences, a large amount of Fe causes spectral overlap on Mn. GFAAS, DRC-ICP-MS and SF-ICP-MS methods show good agreement for Mn in urine and blood across the clinical range studied and, when properly optimized, are suitable for monitoring Mn in blood and urine. In contrast, Q-ICP-MS in standard mode exhibits a positive bias of similar to 1 mu g L(-1) in urine and similar to 4 mu g L(-1) in blood with respect to the other three methods investigated. The implications of this work are important when biomonitoring studies are compared and reference ranges are established.