3.2. Mineral from Feed Ingredient Categories
Table 2 indicates mineral concentrations of feeds by categories. Percentages of NRC requirements of selected minerals fulfilled by each feed category are depicted in
Table 3. It was not possible from the current dataset to discriminate mineral concentrations of individual forages. It is interesting to note the wide variability of mineral requirements satisfied by ingredients and mineral concentrations of ingredients. Indeed, P requirements fulfilled by the forage in the ration varied from 32.7% (percentile 1) to 93.0% (percentile 99), whereas Co requirement fulfillment by forages ranged from 0.0% (percentile 1) to 948.2% (percentile 99). These results highlight the fact that mineral concentrations of ingredients are affected by several factors such as soil type [
4], soil contamination, and sampling method. This could also be the case for homegrown components such as corn grain, other cereals, and soy products. Disregarding mineral interactions that could occur in the rumen and taking into account absorption coefficients, results suggest that, for more than 50% of the participating herds, forages were sufficient to attain the NRC Co, Fe, and Mn requirements. Nevertheless, forage source solely did not suffice to reach the requirement adequacy of P, Cu, and Zn (
Table 3). Sprinkle et al. [
4] also obtained similar results. Nevertheless, very little data are available on the mineral absorption efficiency of basal diet ingredients, which could have a major effect on supply calculations. The huge variability in percentages of mineral requirements fulfilled by commercial energy and protein supplements and minerals and vitamins could be explained by the fact that, in some herds, minerals and vitamins were added in the commercial energy or protein supplement. Hence, in some rations of these herds, no additional mineral and vitamin supplement was added, as observed in
Table 1. In 50% of herds, the mineral supplement alone fulfilled the daily requirements, or was close to these, for Co, Fe, and Mn. It is interesting to note that the mineral supplement satisfied from 0.9% to 47.9% (percentile 1 to 99) of the P requirement, hence not exceeding 100% of needs. This shows that nutritionists pay special attention to this mineral in order to avoid an overall diet excess of P. Median concentrations of Cu and Zn of mineral and vitamin supplements obtained in the current study were similar to what has been reported by Li et al. [
16].
3.3. Recommendations According to DIM
Averaged DIM were 12.5, 50.9, 140.7, and 291.3 ± 1.7 by DIM category ≤21, between 22 and 80, between 81 and 199, and ≥200, respectively (
p < 0.0001;
Table 4). As expected, the milk yield was greater and milk fat and protein concentrations were lower during the lactation peak between 22 and 80 DIM than other DIM categories. Except for Co, for which a non-factorial approach was used to compute requirements, studied mineral recommendations from NRC were greater in the fresh group (<21 DIM) compared with other DIM categories (
p < 0.0001;
Table 4). The same results were obtained for P recommendations from INRA and EAAP. This could partly be explained by lower predicted DMI in those cows. Moreover, those cows had greater fat-corrected milk than cows in mid- and late-lactation, and this implies that they have higher requirements to support lactation. Greater Cu, Fe, Mn, and Zn requirements above 200 DIM compared with between 81 and 199 DIM could be explained by the increased demand for pregnancy. There is no specific recommendation for Fe in INRA tables and a non-factorial approach regarding Co, Cu, Mn, and Zn was adopted by the committee [
32]. A non-factorial approach was also used for these last minerals by the EAAP committee as well as for Fe [
33]. As also outlined by Sinclair and Atkins [
17], requirement dissimilarities exist between recommendation sources. This is the case regarding Mn, where NRC recommendation is well below INRA and EAAP recommendations. It is worth noting that Weiss and Socha [
34] have found that Mn NRC requirements might be underestimated.
Except for Co and Fe, dietary concentrations of P, Cu, Mn, and Zn changed with DIM categories (
p < 0.02;
Table 5), where they were greater <21 DIM than in late lactation. This is in line with NRC recommendations that support the increasing demand for milk production coupled with the limited DMI during this period. Regarding NRC recommendations for P, Cu, Fe, Mn, and Zn, dietary concentrations were closer to the requirements before 21 DIM than thereafter in the lactation (
p < 0.0001;
Table 5). However, as a non-factorial approach was used for studied trace minerals in INRA and EAAP recommendations and as dietary concentrations were greater in early postpartum, trace mineral overfeeding in early lactation was higher than in late lactation. It could be noted that for all trace minerals, regardless of the recommendation sources and DIM categories, the average dietary concentration exceeded the guidelines, as previously observed [
16,
17,
18]. None of these, however, surpassed the maximum tolerable levels for Co, Mn, Fe, and Zn [
35]. Mineral toxicity in animals is quite unusual, as an adaptation mechanism occurs to increase manure excretion according to the increase in supply [
1]. In a review, López-Alonso [
5] have stressed that this current practice of providing more minerals than needed in intensive systems could have detrimental effects on ecosystems.
3.5. Trace Minerals
In the last years, studies have been conducted to evaluate the effect of trace mineral supplementation sources, i.e., either inorganic or organic sources, especially in early lactation, on cow performance, immunity, health, and oxidative metabolism [
13,
14,
38,
39]. Unfortunately, it is not possible from the current assessment to discriminate the source of trace mineral supplement given to the cows. Milk production per cow has increased remarkably over the last years and whether trace metal requirements as per NRC [
1] is sufficient to express optimal performance and metabolism function has been questioned [
40]. Hence, some studies have investigated the effect of feeding greater trace-metal concentrations than the NRC recommendations [
14,
15]. Regardless of the trace-metal sources, dietary Co, Cu, Mn, and Zn concentrations in the study of Osorio et al. [
14] represented percentiles 90, 10, 35, and 38, respectively, of the current diet distribution of the 100 herds for cows below 21 DIM. This means that, although dietary concentrations of Co, Cu, Mn, and Zn were already higher than NRC recommendations in Osorio et al. [
14], 10%, 90%, 65%, and 62% of herds in the current study provided even greater amounts to fresh lactating cows.
Cows do not have a Co requirement per se, but the microorganisms dwelling in their rumen do need Co to synthesize vitamin B
12 [
41], which, in turn, is needed by the cow. Co was fed in excess between 452% and 509%, 102% and 123%, and 507% and 569% relative to the NRC, INRA, and EAAP recommendations, respectively, among DIM categories (
Table 5). Moreover,
Table 6 shows that all herds, if following NRC and EAAP recommendations, fed dairy cows with an excess of Co. In a cross-sectional study involving American and Canadian farms, Duplessis et al. [
42] also found that Co concentrations exceeded the NRC requirement and hence they failed to find a relationship between dietary Co concentration and plasma vitamin B
12 concentration. Co is usually not reported in surveys assessing the difference between dietary mineral concentrations and the requirements.
Dietary Cu concentration was greater in the ≤21 than in the ≥200 DIM category (
p = 0.02;
Table 5) and was similar to what has been reported in Wisconsin and California herds [
16,
18]. Bidewell et al. [
43] reported a case of Cu poisoning for cows receiving a ration having 50 mg/kg DM of Cu. The maximum tolerable level of Cu was set at 40 mg/kg of DM [
35]. One herd was fed a ration with Cu concentration above 40 mg/kg DM in the current study. Copper is the trace mineral having the greatest potential to cause toxicity, as the difference between the requirement and the toxic level is small [
6]. Hence, nutritionists should pay special attention to Cu to avoid overfeeding. The median of the percentage of dietary Cu concentration in excess to the NRC recommendation was 52% and was the closest to the NRC recommendation regarding trace minerals (
Table 6). Dietary Cu absorption is known to decrease with increasing dietary sulfur and molybdenum [
1]. Unfortunately, molybdenum concentration in the diet was not available in the current study. Nevertheless, results suggested that dietary sulfur averaged 0.21% of DM (
Table 1), which is close to the required sulfur concentration [
1].
Dietary Fe concentration did not change according to DIM categories (
p = 0.17) and averaged 226 (SD: 88) mg/kg of DM (
Table 5). Regarding NRC recommendations, as lactation progressed, the percentage of dietary Fe concentration relative to the recommendation progressively increased (
p < 0.0001). No significant effect of DIM categories was observed for the Fe EAAP recommendation regarding the percentage of dietary concentration over the requirements (
p = 0.17). Among the studied minerals, Fe was the most overfed according to the NRC and EAAP recommendations (
Table 5), as also previously observed [
17,
18]. Castillo et al. [
18] explained this result by the fact that forages contain large amounts of Fe, but with low bioavailability, lowering the risk of toxicity for the animal [
35]. Moreover, Fe is rarely intentionally added in the mineral supplement. In the current study, for both NRC and EAAP recommendations, all cows were fed above the requirements as percentile 1 was 469 and 63% relative to the recommendations, respectively (
Table 6).
The dietary concentration of Mn was greater below 21 than above 81 DIM (
p = 0.02;
Table 5). The dietary recommendation of Mn was greater for INRA and EAAP than NRC (
Table 4). This is why the Mn concentration in the diet was closer to the INRA and EAAP than the NRC recommendations (
Table 5). As mentioned above, Weiss and Socha [
34] have found that Mn requirements for lactation cows are about 1.6 higher than the NRC recommendation. As for other trace minerals, Mn was also fed in excess, as also observed by others [
17,
18], and it was different according to the stage of lactation and recommendation sources (
p ≤ 0.02;
Table 5). According to the INRA and EAAP recommendations, some cows were fed below their requirements (
Table 6). Nevertheless, this was not the case according to the NRC recommendations, as the dietary Mn concentration was 107% above the requirements at percentile 1. Manganese toxicity is not a common problem in ruminant, as the maximum tolerable amount is 2000 mg/kg of DM [
1] and no adverse signs were observed when dietary Mn was below this threshold [
35]. In the current study, the highest Mn concentration in the diet reached 285 mg/kg of DM.
The dietary Zn concentration was greater in cows below 21 than above 81 DIM (
p = 0.001;
Table 5). Zinc concentrations in diets observed by Li et al. [
16] in Wisconsin, USA were similar to the current assessment. Nevertheless, surveys conducted in European and in central and northern England dairy farms [
17,
44] have shown that Zn concentrations in the diet were smaller by between 14% and 30% than in the current study, probably caused by the European Union legislation regarding trace minerals [
45]. Cows below 21 DIM were fed closer to their Zn NRC recommendations than other DIM categories (
p < 0.0001), whereas the opposite was obtained regarding INRA and EAAP requirements (
Table 5). About 90% of cows were fed above their Zn requirements, regardless of the source. Along with Cu, Zn is one of the trace minerals fed closest to the NRC requirements (
Table 6), with a median of 65% in excess of the requirements. Sobhanirad et al. [
46] and Sobhanirad and Naserian [
47] did not find adverse effects of feeding rations with Zn concentration greater than 500 mg/kg of DM. In the current study, all cows were fed below the Zn concentration used in Sobhanirad et al. [
46].