Arsenic Determination: A Muddle of Methods

The world’s rice supply contains concentrations of toxic arsenic compounds that could put many consumers at risk of long-term adverse health effects (1). Regulatory and advisory agencies are introducing limits; for example, in 2015, the Commission Regulation for the EU adopted a cut-off of 200 µg kg-1 for inorganic arsenic in white rice and 100 µg kg-1 for rice destined for the production of infant foods. To understand the details of the contamination and how it might be evolving with time – and to enforce the regulations – we need reliable information about the concentrations of the various arsenic compounds in rice, notably arsenite and arsenate.

Reports of studies of arsenic and rice currently appear in the peer-reviewed literature at a rate of about 80 per year, having risen in an almost linear fashion from near-zero in 2001. Reports focusing on analytical methods for arsenic compounds in rice are a little harder to identify, but I would estimate that there are now 10–15 per year. I’m an active researcher in the field and write at least one review per year (2, 3), so I read large numbers of articles in which yet another procedure for the determination of arsenic compounds in rice is described. My conclusions? That there is no convergence on any aspect of the method – and that both writers and reviewers of manuscripts are falling down on the job.

There are several important questions that readers ask: i) in what way is your method an improvement on previously published methods? ii) what, exactly, is new in your procedure? iii) what, precisely, did you do in each step of the method? and iv) why did you choose those particular steps? Many articles fail to answer these basic questions.

Authors describing HPLC-ICP-MS in particular omit a lot of important information. As a start, readers should know whether the chromatography (choice of column, mobile-phase composition, temperature) is new. If not, there should be a citation to the article in which the method was originally described. If the chromatography is new, the researchers should include some explanation as to why this was necessary, together with some description of the optimization experiments and a clear statement of what figures of merit were being optimized. Some chromatographic performance parameters, such as resolution and plate numbers, should be provided and, if gradient elution was chosen, how long re-equilibration took. As isocratic elution is preferred, especially for ICP-MS detection, the decision to develop a method with gradient elution needs to be justified.

A chromatogram of standards – each of which contains the same concentration of arsenic – should be provided, together with a clear description of the calibration procedure. Details of possible compound-dependent responses is vital – if the researchers claim that they observed none, what is the evidence? If separate calibrations were constructed for each component, the slopes of the calibrations should be given. As rice contains enough chlorine to interfere if it co-elutes with an arsenic species, the fate of chloride in the chromatography should be clearly stated (and shown), or the detector operated in a mode that eliminates this problem. It should also be made clear whether the first peak represents chromatography or is the response to compounds running with the solvent front, and a “chromatogram” with an appropriate marker species should be shown. There are similar problems with the descriptions of the sample preparation procedures, but I’ll save those discussions for another article.

I appreciate that journals have limited space, but much of this evidence could be provided in the “supplemental information” file. Without this crucial information, the convergence on methodology that is so sorely needed will remain a distant goal.