Samples

Meat cuts were purchased from local supermarkets and butcheries in order to consider different origins in terms of production system (meat labels) and anatomical location: beef (sirloin, rib-eye and braising steak), pork (neck steak, chop and loin), lamb (chop and loin), chicken (breast with and without skin and leg with skin) and foal (sirloin and filet). Beef, pork and lamb samples were purchased in two series, from June to August 2004 and from October to December 2004. Foal and chicken samples were purchased from January to March 2005. All samples were cut out of the center of the pieces of fresh meat with a ceramic knife to avoid contamination with trace metals. Samples were vacuum packaged, frozen at −18°C and stored until analysis. Two standard reference materials (SRM 1546 Meat Homogenate and SRM 1577b Bovine Liver) from the National Institute of Standards and Technology (NIST) were prepared to test the accuracy of methods.

Sample digestion procedure

In order to determine the mentioned trace elements, a closed quartz vessels and microwave oven (MLS Ethos, MLS-GmbH, Leutkirch, Germany) digestion procedure was applied. One gram of meat sample, and 500 or 250 mg of the certified reference materials SRM 1546 or SRM 1577b, respectively, were accurately weighed into quartz digestion vessels. Two milliliters of hydrogen peroxide (H₂O₂ 30% pro analysi, Merck, Darmstadt, Germany) and 8 ml nitric acid 65% (HNO₃ suprapur, Merck, Darmstadt, Germany) were added to each tube, which was then closed. For the spiking test, an appropriate amount of all measured elements (Mn, Cu, Mo, Cd, Pb, Fe and Se) was also added. Digestion was conducted by applying a four-step program as follows: 250 W for 4 min, then 3 min at 350 W followed by 450 W for 3 min and finally 3 min at 500 W. After the vessels had cooled down, the digests were transferred into 20 ml volumetric flasks and filled to the mark using purity water (obtained from a Milli-Q water purification system, Millipore, France). Prior to analysis, an aliquot of these samples was spiked with the internal standard solutions and diluted further by a factor of two. Within each digestion run, two samples were randomly chosen for quality control measurements. One sample was digested in duplicate to check for reproducibility of the digestion and analysis, while the other sample was spiked with known concentrations of the elements measured before digestion to monitor potential analyte losses during digestion and further sample preparation. Furthermore, one digestion blank was prepared by the same procedure, which contained only the chemicals used for digestion without a sample. The reference materials NIST SRM 1577d (bovine liver, certified values for Mn, Fe, Cu, Zn, Se, Mo, Cd and Pb) and 1546 (meat homogenate, certified values for Mn, Cu and Zn) were used to verify the accuracy of the digestion and calibration protocol.

Determination of manganese, copper, molybdenum, cadmium and lead concentration by inductively coupled plasma mass spectrometry (ICPMS)

This suite of elements was analyzed in an early session using an Agilent 7500cs ICPMS (Agilent Technologies, Waldbronn, Germany) in normal operation mode. For sample introduction, the instrument was equipped with a quartz concentric nebulizer and spray chamber. The solutions were introduced to the nebulizer by a peristaltic pump at a flow rate of 0.7 ml/min. The operating parameters of the instrument were adjusted daily for optimum sensitivity, while keeping the formation of refractory oxides formed within the ICP to a minimum (that is CeO⁺/Ce⁺<0.5%) to avoid excessive spectral interferences. Typical operating conditions are listed in Table 1. All element concentrations were determined against external calibration using synthetic acidic multielement calibration standards. The accuracy of the calibration was assessed by analyzing the certified reference materials NIST SRM 1577 and 1546. In all cases, the measured concentrations in the reference materials were within 10% of the certified values. Repeatability of the duplicate samples was between 2% for concentrations higher than three times the limit of quantification (concentration equivalent to 10 times the standard deviation of the blank measurements) and 50% for samples that were close to the limit of detection. Spike recoveries typically varied between 88% and 100%. The limits of quantification were as follows: Mn 0.05 mg/kg, Cu 0.1 mg/kg, Zn 0.5 mg/kg, Mo 0.005 mg/kg, Cd 0.001 mg/kg and Pb 0.001 mg/kg (note that instrumental limits of detection were at least an order of magnitude lower). To compensate for nonspectral matrix effects, two internal standards were employed. Rh was used for Mn, Cu, Zn, Mo and Cd and Lu for Pb.

Table 1

Operating condition for the Agilent 7500 cs ICPMS

Determination of selenium and iron by dynamic reaction cell inductively coupled plasma mass spectrometry (ICP-DRC-MS)

Owing to a high abundance of background ions from the ICP on the major isotopes of Se and Fe (Ar₂⁺ and ArO⁺, respectively), these elements cannot be determined interference free by standard ICPMS. Thus, the trace and ultratrace determinations require suppression of these interferences. Dynamic reaction cell ICPMS allows the attenuation or even removal of these types of spectral interferences by selective ion molecule reactions before the ion population generated in the ICP is mass analyzed. In this study, an Elan 6100 DRC^plus (PE/Sciex, Concord, Ontario) was used for the determination of Se and Fe. The instrument utilizes a quadrupole ion guide in a reaction chamber, where interferences and analytes may undergo reactive collisions with a gas. The quadrupole ion guide may furthermore be adjusted to reject ions from a specific m/z range in order to avoid the formation of reaction products that may interfere with the analyte of choice. Methane, as the reaction gas, allows the suppression of the dominating Ar₂⁺ and ArO⁺ interferences by six orders of magnitude through a charge exchange reaction, which does not occur for Se and Fe at a similar rate (Tanner et al., Reference Tanner, Baranov and Bandura2002; Hattendorf and Günther, Reference Hattendorf and Günther2003). The occurrence of potentially formed new polyatomic ions inside the dynamic reaction cell was tested and the formation of the adduct ion ⁶⁴Zn(CH₄)⁺ was found to occur. This new interference on the isotope ⁸⁰Se, however, could be eliminated by adjusting the transmission properties of the dynamic reaction cell (i.e. the parameter RpQ). All other operating conditions were optimized to yield the highest signal/background ratio, especially for ⁸⁰Se⁺, while minimizing the instrumental background. The same suite of samples as in the first sequence was analyzed using an identical quality control scheme, except for not analyzing NIST SRM 1546 that is not certified for Fe and Se. Repeatability for the duplicate samples was better than 8% in all cases. The limits of quantification were as follows: Fe = 0.1 μg/kg and Se = 40 μg/kg. The operating conditions are listed in Table 2.

Table 2

Operating condition for the Elan 6100DRC^plus ICP-DRC-MS

Cadmium and lead

The results show a high variability (CV = 58% to 127%) in cadmium concentrations for foal sirloin and filet as well as for lamb chop, which also has the highest cadmium value (Table 3). The remaining meat cuts had a mean cadmium concentration between 0.5 and 0.7 μg/100 g and lower CVs (2% to 34%). In comparison, the analyzed lead concentrations were more consistent and varied between 1.8 and 2.0 μg/100 g with CVs ranging from 3% to 35%. González-Weller et al. (Reference González-Weller, Karlsson, Caballero, Hernández, Gutiérrez, González-Iglesias, Marino and Hardisson2006) found a much larger variability for cadmium and lead. They suggested that the variability in their samples is considered normal since possible sources of this metal are numerous. However, they also indicated that lead and cadmium concentrations depend on environmental conditions and food production methods. Compared with the results of González-Weller et al. (Reference González-Weller, Karlsson, Caballero, Hernández, Gutiérrez, González-Iglesias, Marino and Hardisson2006), the mean concentrations for cadmium and lead in this study are higher, except for cadmium concentration in pork where the results are similar. Compared with other studies, the analyzed lead concentrations are comparable to those measured in Finland and Slovenia (Niemi et al., Reference Niemi, Venalainen, Hirvi, Hirn and Karppanen1991; Doganoc, Reference Doganoc1996), but were higher then those reported by Tahvonen and Kumpulainen (Reference Tahvonen and Kumpulainen1994) and Hecht and Kumpulainen (Reference Hecht and Kumpulainen1995) for beef and pork. Lead concentrations in horse meat were found to be lower by a factor of 0.5 compared with those determined by Hecht and Kumpulainen (Reference Hecht and Kumpulainen1995). Cadmium concentrations are higher compared with the last mentioned study but again with the exception of horse meat, which is found to be much lower in our study. The maximum lead and cadmium levels allowable in meat from different species are 100 and 50 μg/kg (EU Directive 466/2001). The analyzed concentrations of these metals in this work are all below these limits.

Iron and zinc

The iron concentration varied greatly among the species in this study (Table 4). Among the red meat, foal meat was characterized by the highest iron content. Lamb and beef showed a similar range of iron concentrations, from 1.6 mg/100 g determined in beef (sirloin US) to 2.6 mg/100 g determined in lamb (loin). Markedly lower iron concentrations were determined in pork and chicken meats. The iron concentrations are in agreement with Hecht and Kumpulainen (Reference Hecht and Kumpulainen1995), Chan et al. (Reference Chan, Brown, Lee and Buss1995) and Leonhardt and Wenk (Reference Leonhardt and Wenk1997). The last reference reported lower CVs (10% to 35%) compared with CVs for iron concentration in this study (15% to 93%). The main reason for the variability in the examined trace element content and its CVs might be that different breeds and feeding system arose in Switzerland in the last few years to comply with the provisions of a variety of meat labels that were established. Similar to our results, Schricker et al. (Reference Schricker, Miller and Stouffer1982a) and Carpenter and Clark (Reference Carpenter and Clark1995) found that the iron concentrations varied between different cuts from the same species.

Therefore, it is important to draw interspecies comparisons with the same cuts or muscles.

The zinc values showed that concentrations of this element vary even more than those of iron between the different muscles. Comparable findings are reported by several studies (Schricker et al., Reference Schricker, Miller and Stouffer1982b; Marchello et al., Reference Marchello, Slanger and Milne1985; Hecht and Kumpulainen, Reference Hecht and Kumpulainen1995; Leonhardt and Wenk, Reference Leonhardt and Wenk1997). Cassens et al. (Reference Cassens, Briskey and Hoekstra1963) reported that the zinc content in various porcine muscles varied with color and myoglobin concentrations and that dark muscles had greater concentrations of zinc than light ones; they also found increased concentrations of zinc in more active muscles. The measured values for zinc are similar to values reported by Leonhardt and Wenk (Reference Leonhardt and Wenk1997) and Hecht and Kumpulainen (Reference Hecht and Kumpulainen1995) except for chicken breast, which was found to be lower than the results reported by the Hecht and Kumpulainen study, but in agreement with the results of Chan et al. (Reference Chan, Brown, Lee and Buss1995). The calculated CVs (4% to 45%) for zinc were the lowest in this study but higher compared with the value (7% to 18%) found by Leonhardt and Wenk (Reference Leonhardt and Wenk1997). Pennington et al. (Reference Pennington, Schoen, Salmon, Young, Johnson and Marts1995) reported CVs in the range of 10% to 24% for zinc in different cooked meat cuts. Only two CVs for zinc in this study were above this value, foal sirloin (39%) and chicken leg with skin (45%). A possible explanation for the mainly low CVs of zinc could be that the tissue concentration of this element is primarily genetically determined and is only slightly influenced by feed composition (Flachowsky and Jahreis, Reference Flachowsky and Jahreis1995).

Selenium

Se is not considered essential for plant growth, and plants do not control Se uptake (Ellis and Salt, Reference Ellis and Salt2003). Consequently, Se concentrations in plants generally reflect the concentration and availability of Se in soils (Combs, Reference Combs2000; Hintze et al., Reference Hintze, Lardy, Marchello and Finley2001), which vary greatly depending on the presence of underlying Se-containing geologic formations (Kabata-Pendias, Reference Kabata-Pendias1998). Therefore, animals raised using low-Se feedstuffs deposit relatively low concentrations of this element in their tissues, while animals raised on a diet high in Se yield food products with a higher Se concentration. Due to the needs of livestock for Se to prevent debilitating deficiency syndromes, Se (still usually in the form of sodium selenite) is commonly used as a feed supplement in commercial animal agriculture in many parts of the world (Combs, Reference Combs2001). This practice became widespread in Switzerland within the last 25 years and thereby meat has become a constant and an important Se source. The mean Se concentration, reported in Table 5, was higher than reported by Haldimann et al. (Reference Haldimann, Dufosse, Mompart and Zimmerli1999). This indicates that Se as a feed supplement is used more extensively than 6 years before. When comparing the Se concentrations with data of established food composition tables, similar results are reported by Chan et al. (Reference Chan, Brown, Lee and Buss1995) for beef, pork and chicken. Souci et al. (Reference Souci, Fachmann and Kraut2000) recorded lower Se concentrations for beef, foal and lamb but comparable concentrations for chicken. US Beef had the highest selenium concentration (30 and 44 μg/100 g), around three to four times higher than Swiss beef. The main reason is the higher Se concentration in soil combined with the practice of transferring beef cattle to feedlots before slaughter, and while there, feeding them a defined diet supplemented with Se (Keck and Finley, Reference Keck and Finley2006). The CV for this element recorded in this study ranged from 13% to 82% within species and cut due to the influence of feedstuff and supplements. For beef, the CVs for Se concentration vary between 65% and 82% in the Swiss cuts compared with the US beef with CVs of 20%. In the last few years, organically raised beef increased in popularity in Switzerland. As these animals do not get any mineral supplements, the Se concentration will depend on the natural supply and therefore on the availability according to the geographic origin.

Manganese, copper and molybdenum

The measured manganese concentrations are lower than the values reported by Souci et al. (Reference Souci, Fachmann and Kraut2000) but also show, as well, a high unsystematic variation between animal species and different cuts (Table 6). On the contrary, Chan et al. (Reference Chan, Brown, Lee and Buss1995) reported mean values for Mn, which have a similar order of magnitude but show a lower variation. However, there are also studies that show similar Mn concentrations and variations (Hecht and Kumpulainen, Reference Hecht and Kumpulainen1995). The high variability (13% to 142%) for this element indicates that Mn is affected, for example, by different feeding systems. On the other hand, homeostatic mechanisms maintain most tissue manganese concentration within certain limits (National Research Council, 2000).

The copper concentrations of chicken meat differed markedly from the other meat cuts analyzed, in which much higher concentrations are found. The concentrations in meat cuts from lamb, beef and pork are in good agreement with the values reported by Hecht and Kumpulainen (Reference Hecht and Kumpulainen1995) and Lombardi-Boccia et al. (Reference Lombardi-Boccia, Lanzi and Aguzzi2005). However, concentrations in horse meat reported by Hecht and Kumpulainen (Reference Hecht and Kumpulainen1995) are twice as high compared with the study of Lombardi-Boccia et al. (Reference Lombardi-Boccia, Lanzi and Aguzzi2005) and with ours. In the food composition table of Souci et al. (Reference Souci, Fachmann and Kraut2000), Cu concentrations are six to ten times higher and differ between muscles to a greater extent than in the present study. The within-CV of the various cuts, which is not provided by the food composition tables cited, differed greatly (13% to 224%), with the highest variability for the chicken meat. As with zinc and iron, copper concentrations in meat can vary in the different types of muscle. In cattle, it has been demonstrated that copper concentrations in muscle are inversely related to muscle lipid concentrations (Langlands et al., Reference Langlands, Donald and Smith1987).

Molybdenum concentrations in meat are very interesting since these results show that chicken meat (2.4 to 3.2 μg/100 g) contains around twice as much as mammalian meat. This is similar to the study of Hecht and Kumpulainen (Reference Hecht and Kumpulainen1995) who found even greater differences due to lower Mo concentrations in beef and pork meat and higher values for chicken meat, compared with this study. Only a few figures for Mo concentrations can be found in the food composition table of Souci et al. (Reference Souci, Fachmann and Kraut2000) for meat and they are up to ten times higher than reported by Hecht and Kumpulainen (Reference Hecht and Kumpulainen1995) and found in the present study. In other food composition tables, Mo is not listed. The CV of molybdenum ranged from 2% to 63% and was therefore the third lowest of all trace elements analyzed in this study.

Essentiality v. toxicity

Essential trace elements are those compounds that need to be present in the human diet to maintain normal physiological functions. Risk assessment of trace elements has examined two ends of the toxicity-deficiency spectrum: that associated with intakes that are too high and results in toxicity and that associated with intakes that are too low and results in nutritional deficiency problems (Goldhaber, Reference Goldhaber2003). Meat is known to be a source of trace elements but has, as well, been discussed in terms of accumulating heavy metals such as cadmium and lead. Several studies have been carried out with respect to the contribution of meat consumption towards fulfilling the requirements for several essential trace elements (Leonhardt et al., Reference Leonhardt, Kreuzer and Wenk1997; Haldimann et al., Reference Haldimann, Dufosse, Mompart and Zimmerli1999; Chanson et al., Reference Chanson, Brachet, Grolier and Rock2003; Lombardi-Boccia et al., Reference Lombardi-Boccia, Lanzi and Aguzzi2005). While toxic concentrations of essential elements in muscle are generally rare, offal, such as liver and kidney, often accumulate higher metal concentrations. The US Food and Drug Administration (FDA) has defined a tolerable upper intake level (UL) for the essential trace elements. The UL is the maximum level of total chronic daily intake of a nutrient (from all sources) judged to be unlikely to pose a risk of adverse health effects to humans (Scientific Committee on Food, 2006). Concentrations in the meat cuts analyzed in this study are far below the ULs for all trace elements. The results suggest that, at present, there is no need to recommend restrictions for the human consumption of meat marketed in Switzerland.