Analysis of soluble or titanium dioxide derived titanium levels in human whole blood: consensus from an inter-laboratory comparison.

No Thumbnail Available
Change log
Koller, D 
Bramhall, P 
Devoy, J 
Goenaga-Infante, H 
Harrington, CF 

Exposure to titanium (Ti), via the ingestion of pigment grade Ti dioxide (TiO2), is commonplace for westernised populations. It may also occur as a consequence of metal ion leaching in subjects bearing Ti-containing implants. Accurate exposure analysis requires fit-for-purpose analytical methodology, especially for true measures of baseline levels. Inductively coupled plasma (ICP) techniques are, mainly, now used for bio-analysis of Ti. Since whole blood reference materials, certified for natural low levels of Ti, are not currently available, we undertook an inter-laboratory comparison of pooled human blood from fasted volunteers ±low level (+∼2.5 μg L-1) or high level (+10-20 μg L-1) spikes of soluble Ti or TiO2 particles. Seven established laboratories were enrolled to analyse the samples using ICP based techniques, which included at least one of ICP optical emission spectrometry (ICP-OES), high resolution ICP mass spectrometry (HR-ICP-MS), triple quadrupole ICP-MS (ICP-MS/MS) or single quadrupole ICP-MS (SQ-ICP-MS). Five laboratories diluted the blood for analysis whilst two performed acid digestion. Overall, we showed that the laboratories could, mostly, quantitatively detect modest levels of spiked Ti in blood. Markedly varying levels of Ti, however, were reported for the same baseline pooled sample (0.4-24.6 μg L-1) and, in this study, specificity was poor for SQ-ICP-MS. Digestion of samples caused sample contamination compromising limits of detection and accuracy, whilst simple dilution had no such problem, and remained linear in response for spikes with ionic and TiO2 particles. We conclude that measuring baseline levels of Ti in whole blood is challenging but should be readily achievable down to 0.5-1.5 μg L-1, if sample preparation avoids contamination and instrument techniques are used that negate polyatomic or isobaric interferences from the sample matrix. We also remind those relying upon Ti bio-analytical data for their experimental outcomes that (a) spiking and recovery experiments provide information only on linearity of detection but not at all on accuracy as this will not detect constant positive errors and that (b) biological standard materials for Ti generally contain high levels of the analyte and tend to mask baseline analytical errors. Caution may be required in interpreting the findings of some published Ti/TiO2 bio-exposure studies.

Consensus, Humans, Limit of Detection, Spectrophotometry, Atomic, Tandem Mass Spectrometry, Titanium
Journal Title
Conference Name
Journal ISSN
Volume Title
Royal Society of Chemistry (RSC)
Medical Research Council (MR/R005699/1)