Development and validation of an analytical method for the determination of arsenic, cadmium and lead content in powdered infant formula by means of quadrupole Inductively Coupled Plasma Mass Spectrometry

Cadmium (Cd), lead (Pb) and arsenic (As) can enter the food chain through the environment and/or as a consequence of the manufacturing process making foodstuffs the main human exposure route to these chemical elements. The risk associated with this exposure is of such a big concern for human health that the European Food Safety Agency recommends to reduce the exposure to Cd and Pb so as to protect especially vulnerable subgroups of population (e.g., infants). Therefore, the setting of new maximum levels (MLs) for chemical elements in infant formulae (e.g., for Cd) or the reconsideration of the existing ML for Pb is under discussion. On this basis, the availability of analytical methods, precise, accurate and sensitive enough to quantify low concentration values, is a key point especially for official control laboratories that have to state the sample compliance using a fully validated method with an associated uncertainty compliant with the requirements specified in the pertinent regulations. This work describes the development and validation of an analytical method to quantify As, Cd and Pb in powdered infant formulae based on animal protein at values of concentration close to the MLs that are likely to be set. The results obtained make the method suitable for a precise and accurate determination of these chemical elements at these low concentration values. In particular, the results for limit of quantification (LoQ) were respectively (μg kg−1): As 6.2, Cd 1.2 and Pb 4.5. While for the recovery rates the following percentages were obtained: As 105%, Cd 98% and Pb 108%. The expanded uncertainties were found extremely satisfactory (Cd 13% and Pb 19%). The LoQ and the uncertainty for Pb meet the requirements set in Commission Regulation (EC) No. 333/2007 and following amendments being lower than the maximum values allowed. Even for Cd the expanded uncertainty resulted adequate in relation with the low concentration considered.