The mineral concentrations in paired whole blood and serum samples (n = 10) are summarised in Figure 1. Iron was the most abundant element in whole blood (364 ± 64 mg L⁻¹, ranging from 211 to 381 mg L⁻¹), in concentrations approximately 300 times higher than those in serum. Whole blood Fe is mainly represented by Hb, which contains four Fe atoms per molecule [1]. The variations in the ratio of Fe in whole blood:serum can almost exclusively be attributed to the number and volume of erythrocytes. Physiological variations in red blood cells are observed among individuals, as well as in numerous pathological conditions, particularly those causing anaemia. Overall, the results obtained indicate that haemolysis can strongly influence the determination of the Fe status of cattle and could lead to the overestimation of Fe, although this error can potentially be corrected if the HD is calculated by considering that 1 g of haemoglobin contains 3.47 mg of Fe [29].

Chromium, Zn and Se concentrations were also significantly higher in whole blood than in serum, with whole blood:serum ratios of 3.73, 2.71 and 1.43, respectively. For all of these elements, the largest interindividual variations were found for the whole blood samples (Figure 1). Both Se and Zn are cofactors or essential intra erythrocyte enzymes that are closely involved in the immune system, which could explain the large individual variation within subjects. Selenium is associated with GPX1, the most abundant glutathione peroxidase in the body, and is also responsible for most of the blood Se concentration (erythrocyte Se usually represents 60–73% of the Se in whole blood in cattle [1,22]). Zinc is also essential for multiple intraerythrocytic enzymes such as superoxide dismutase, carbonic anhydrase and lactate dehydrogenase (LDH) [1,3]. The high concentration of LDH in red blood cells leads to its overestimation when measured in serum with HD values lower than 0.5 g Hb L⁻¹ [23]. Regarding Cr, when present as Cr (III), it constitutes an essential mineral with beneficial influence on immune and antioxidative activity and energy metabolism [30,31] and is transported in serum due to its inability to penetrate red blood membranes [32]. By contrast, the hexavalent form Cr (VI) is highly toxic and carcinogenic [33] and has a high affinity for erythrocytes, where it is reduced to Cr (III) and bound to the β-chain of Hb [32]. In the present study, the higher Cr concentration in the whole blood samples than in the serum probably reflected a slight exposure of the subjects to environmental Cr (VI).

The whole blood concentrations of Mg, Mo and Ca were significantly lower than in serum (whole blood:serum ratios of 0.79, 0.65 and 0.66 respectively). The intraerythrocytic concentrations of these elements are low and, as they are mainly transported bound to plasma proteins in the blood, sample haemolysis would lead to underestimation of the concentrations in serum due to a dilution effect. The Mg and Mo concentrations showed similar inter-subject variability in whole blood and in serum, making it difficult to precisely predict its influence. However, the Ca concentrations in whole blood revealed very low inter-individual variation, suggesting that the effect of haemolysis could potentially be estimated by the analysis of haemolyzed serum. For the other elements, there were no significant differences between the concentrations in whole blood and serum, suggesting that haemolysis would have a negligible effect on the concentrations measured in serum.

The essential role of mineral elements in animal health and production, together with the high prevalence of mineral imbalances worldwide, has led to a large body of research on mineral element metabolism in the last few decades, including the establishment of precise reference values in serum and plasma (for reviews, see [1,16,17]). However, information of mineral element concentrations in whole blood is scarce, with the exception of Se. This is probably because Se deficiency is one of the most prevalent and costly trace element imbalances in livestock and the determination of Se in whole blood provides accurate, long-term information about the Se status of the animal [1,13]. Information about the effect of haemolysis on the determination of mineral status in cattle and other livestock species is also very limited. In a comprehensive review by Herdt and Hoff [17] on the use of blood to evaluate the mineral status in ruminants, apart from Fe, possible leakage from erythrocytes is only mentioned in relation to Zn and Mn. According to these authors, haemolysis or prolonged contact of the serum with the clot leads to the escape of Zn and Mn from red blood cells into the serum, thus producing a false increase in serum concentrations. However, these suggestions are based on studies in humans, in which the whole blood:serum ratios are different from those in bovine species. In humans, concentrations of both Mn [34] and Zn [24] in erythrocytes are 10–20 times higher than in plasma or serum, i.e., a ratio whole blood:serum in the range 5–10. In this study, whole blood Zn concentrations were around three times higher than in serum and no difference was observed for Mn. These results indicate that haemolysis may have less influence on the determination of Zn and particularly Mn in cattle than in humans. For all these reasons, it is important to known how and the degree to which haemolysis affects serum concentrations of minerals.