Important trace element concentrations in ovine liver as determined by energy dispersive handheld X-ray fluorescence spectrometry

Abstract

Trace elements are involved in a variety of biochemical processes essential to life and are required in minute amounts. There are no data available on the use of handheld X-ray fluorescence (XRF) spectrometry to determine concentrations of important trace elements in ovine livers. The aim of this study was to ascertain if the handheld X-ray fluorescence spectrometer will provide reliable concentrations of certain essential trace elements in the livers of sheep. Sheep livers (n=30) were obtained from abattoirs. Wet liver samples taken from 30 liver specimens were blended until homogeneity was achieved. An aliquot of the homogenised liver samples were oven dried at 50°C until a constant mass and were then pulverised using a mortar and pestle to obtain a fine powder. In addition, homogenised liver samples (n = 30) were also submitted for dry ashing. All the prepared liver samples (i.e. wet blended, oven dried and dry ashed) were then analysed using a handheld X-ray fluorescence spectrometer to determine concentrations of copper (Cu), iron (Fe), manganese (Mn), molybdenum (Mo), selenium (Se) and zinc (Zn). A reference laboratory analysed the same liver samples using ICP-MS to determine the concentrations of the above mentioned trace elements (control). The means (mg/kg) of the ICP-MS results on a dry matter basis were: Cu (505), Fe (351), Mn (12.3), Mo (3.8), Se (1.8) and Zn (168). The means (mg/kg) of the XRF oven-dried results were: Cu (502), Fe (289), Mn (11.7), Mo (1.6) and Zn (141.9). Selenium could not be detected in oven-dried samples when using the XRF. The intra-sample coefficients of variation were similar between ICP-MS and XRF for oven dried samples for Cu, Fe and Zn and are within the same order of magnitude for all elements in dry ashed samples when comparing ICP-MS to XRF. However, the intra-sample coefficients of variation for Mn and Mo were approximately an order of magnitude larger using XRF. Although the precision for Se appears to be good when using XRF on dry ashed samples, Se was only detected in a few samples, so this value is not representative of the overall precision of XRF using the dry ashed preparation procedure for Se determination. Selenium was not detectable using XRF on wet blended and oven dried samples. The intra-sample coefficient of variation for Se was relatively high using ICP-MS, suggesting that even the current ‘gold standard’ in detecting trace-elements may be imprecise in measuring Se. Overall, this suggests that the precision of sampling using XRF is relatively good for only Cu, Fe and Zn and relatively poor for Mn and Mo. Furthermore, XRF cannot be reliably used for measuring Se. Bayesian correlation were used to determine the best correlation between XRF and ICP-MS data. Bayesian correlation results are summarised by the median sample Pearson product-moment correlation coefficient (r), the 95% lower (LHPDI) and upper (UHPDI) highest posterior density intervals, the square of the sample correlation coefficient (r2), and the probability that the correlation coefficient is positive. Overall, the oven-dried preparation procedure for XRF appeared to provide the best correlation with the ICP-MS data. For Cu and Zn these correlations were strong and the XRF method may represent a suitable substitute for ICP-MS. For Mn and Fe the correlations were moderately strong and the XRF method may be suitable depending upon the intended application. For Mo the correlation was moderate and XRF cannot be recommended. For Se no XRF method was suitable. The advantage of handheld X-ray spectrometry is that the turnaround time of samples is reduced a great deal. Instead of submitting samples to a laboratory and waiting for results, samples can be analysed more rapidly with the use of a handheld X-ray fluorescence spectrometer.