Inorganic contaminants and composition analysis of commercial feed grade mineral compounds available in Costa Rica

The typical concentrations of essential and contaminant heavy metals in common mineral
ingredients used in feed manufacturing is offered. Some information regarding heavy
metals in compound feed is presented as well.

From all the mineral sources analyzed, without exception, sulfate ion-containing minerals
exhibit a significantly (p??0.001) higher concentration of water (from 42.5 to 71.3 g/100 g dry matter) compared
to other sources. This is an anticipated result as SO₄^2?, in accordance with Hofmeister series (dos Santos et al. 2010]), is a small and multiply-charged ion, with high charge density, a poor disruptive
of water’s molecular binding (kosmotropic) (Wei et al. 2005]) and thus exhibit stronger interactions with water (Plumridge et al. 2000]). This said, is to be concluded that for all practical purposes in these type of
salts, ions are to be treated as hydrated.

The case of calcium phosphates is singular. Degradation of orthophosphate into unavailable
phosphate forms during manufacture (Hoffmann et al. 2011]) is a possibility in those minerals which demonstrate lower values of phosphorus
but calcium values within what is expected for these type of products. This occurs
in exactly two (n?=?2) samples of dicalcium phosphate (Table 1).

Table 1. Mineral contents in typical mineral imported supplements employed in feedingstuff
manufacture

Due to the diverse chemical composition of both feed phosphate sources analyzed, large
differences do exist in their phosphorus availability for different animal species
(Fernandes et al. 1999]; Viljoen 2001]; Petersen et al. 2011]). Phosphates from the same source and produced by the same process, using raw substances
of similar quality should offer consistent phosphorus concentrations. In this respect,
we suggest that chemical data should always be accompanied by bioavailability assays.

In a similar fashion, we have found, for the colorimetric assay of phosphorus, standard
amounts of molybdovanadate to be inadequate after wet acid digestion, probably due
an incomplete complexation of phosphorus by this reagent (data not shown). Therefore,
we suggest using a stoichiometric excess of color-forming agent or another quantitative
technique such as GFAAS or ICP-MS for total phosphorus quantitation.

Another distinctive feature of feed phosphates is that fluorine can be used as a quality
control for the inorganic synthesis and inputs used during its manufacture (e.g. wet
acid route). In this case, only 4.5 % (n?=?1/22) of the samples, exceeded the 2 g/kg guideline for this anion (Table 1). So far, only dicalcium phosphate samples have exceeded regulatory guidelines for
fluorine (Table 1). However, no significant differences (p??0.05) were found for this nutrient when both types of phosphate sources were compared.
Synthesis-wise new technologies for the production of feed phosphates have been introduced
with the sole purpose of reducing this contaminant to acceptable levels (Hoffmann
et al. 2011]).

On another hand, in the case of macronutrients, a worrisome situation was evidenced
by the detection of some inconsistencies in labeling (either by the mineral manufacturer
or in situ), this issue seems to arise more frequently with selenium mineral sources. For example,
samples branded as Microgranâ„¢ Se and sodium selenite were in some cases inverted i.e.
the latter showed low sodium and selenium concentrations (?5.0-10.0 g/100 g) and viceversa;
a reagent grade NaSeO₃ with 99 % purity, when analyzed, should return values for sodium and selenium of
at least 13.1 and 45.2 g/100 g, respectively (data not shown). A couple of the recollected
samples, considered as manganese (IV) oxide, exhibited only concentrations ranging
on the low side of the mg/kg of Mn whereas one might expect for a reagent grade MnO₂ at least 62.6 g/100 g of the metal (data not shown). Yet another example lies within
samples labeled as zinc oxide (n?=?2) in which Zn analysis verified concentrations??90 g/100 g of this metal which
is, chemically speaking, an impossibility in terms of formula, this suggest these
samples were not metal oxides but other species entirely. The same occurs in a sample
(n?=?1) of CoCO₃ (82.4 g Co/100 g) and a sample (n?=?1) of KCl (97.58 g K/100 g) (Table 1).

Some samples exhibit mean values significantly (p??0.001) beneath the guaranteed analysis, important examples include KCl for potassium,
MgO for magnesium, EDDI for iodine, ferrous and calcium carbonate for iron and carbonate,
respectively (Table 1). Congruently, 16.8 % a total of the samples (n?=?73) were found to be in this condition. Incidence of some relevant samples was
as follow: 83.6 (n?=?5), 53.8 (n?=?7), 52.9 (n?=?9), 47.1 (n?=?8), 35.0 (n?=?7) and 28.6 % (n?=?6) for EDDI, CoCO₃, KCl, ZnSO₄, MnO and MgO respectively (Table 1). Other minerals do not differ significantly or are between manufacturer assured
ranges. Only for minerals Microgranâ„¢ Co (n?=?14) and MgSO₄ (n?=?17) all samples were within the expected concentration range (Table 1). In this regard, mineral deficiencies should be circumvented as these nutrients
are relevant to health and their monitoring would consequently could prevent or manage
mineral-associated deficiency diseases (Soetan et al. 2010]). This is of foremost importance especially in countries when said minerals may be
scarce or marginal due to geochemical characteristics of the region. Also, interrelationships
and interferences among the mineral elements should be considered when mineral premixes
are formulated (Soetan et al. 2010]).

Regarding heavy metal concentrations, several mineral samples contravene current legislation
in at least one metal. Number of samples in this condition correspond to 0.5 (n?=?2), 13.8 (n?=?60), 4.1 (n?=?18) and 2.5 % (n?=?11) for As, Hg, Pb and Cd, respectively (Table 2). Overall, 21.1 % (n?=?92) of the samples exhibited concentrations of heavy metals and fluorine above
those stipulated by European guidelines. Values for As, Hg, Pb and Cd ranged from
61.2 mg/kg (MnO) to 1.3 ?g/kg (MgO), 2.6?×?10³Â ?g/kg (MnO) to 1.2 ?g/kg (NaCl), 706.2 mg/kg [Ca(H₂PO₄)₂] to 5.5?×?10^?2 mg/kg (KCl) and 9.46 mg/kg (NaCl) to 3.4?×?10^?1 ?g/kg (KCl), respectively (Table 2). Other especially elevated Hg and Pb concentrations (over the 200 ?g/kg and 100 mg/kg
respective permitted levels) found during our survey are 1439 (FeSO₄), 882 (ZnO) and 603 (Fe₂O₃) ?g/kg and 628.1 (MgO) and 455 (CuO) mg/kg, respectively (Table 2).

Table 2. Contaminant concentrations in typical mineral imported supplements employed in feedingstuff
manufacture

In decreasing order of heavy metal overall concentrations the following minerals showed
the less contamination: MnSO₄??EDDI??KCl (Table 2). Microgranâ„¢ Se, Co and I samples also showed relatively lower concentrations of
heavy metals. This may be expected as this ingredient’s presentation mineral input
is from 1 to 10 g/100 g, maximum. The fact that the one iodine source from organic
synthesis has a lower concentration of heavy metals, sustains our hypothesis that
higher contents come from inorganic synthesis raw materials. In the light of this
findings other organic sources could be examined, especially since iodine organic
salts have demonstrated potential as additives. Actually, iodine consumption through
feed (pet and cattle mostly) is primarily as the compound EDDI (Lyday 2005]).

Of the heavy metals assayed As showed, in general, the lowest values. These results
are especially reassuring considering that all livestock species are susceptible to
toxic effects of inorganic arsenic and some feeds may even contain organoarsenical
species (e.g. roxarsone) as growth promoters to improve feed efficiency (Chapman and
Johnson 2002]), occasionally in combination with ionophores. Strikingly, As has been suggested
to possess some essential or beneficial functions at ultra-trace concentrations (Uthus
2003]).

There is considerably less incidence of relatively elevated As concentrations, however
As regulation is in some cases 100 fold more permissive with respect to Hg. In contrast,
the higher number of incidents of irregular concentrations for Hg is evident. This
is expected as Hg is the most abundant naturally-occurring heavy metal and is emitted
primarily due industrial sources and mining ore deposits (Goyer 1996]). As most information on mercury residues in feedstuffs, data presented here is given
as total mercury concentrations. In this regard, although inorganic mercury toxicity
profile due to accumulation include kidney damage, methylmercury (CH₃-Hg) is the form considered of greatest toxicological concern which very well be non-existent,
considering the nature of the samples tested.

Moreover, up to 30.0 % (Table 2) of samples of sulfur assayed (n?=?20) showed residues of Hg??200 ?g/kg. This result may be explained by sulfur chemisorption
capability of Hg (Feng et al. 2006]). In turn, the same samples showed relatively low concentrations of the other heavy
metals despite its tendency to associate with them.

Overall, calcium phosphates and metal oxides showed a significantly higher (p??0.001) levels of arsenic and lead relative to other mineral sources (Table 2). This fact could be explained due to the fact that minerals such as Zn, Mn and Fe
oxides have such redox potentials that can oxidize As and will thus alter the extent
of As retention. In fact, Mn³⁺/Mn⁵⁺ oxides are strong oxidants that can oxidize and sequester many trace metals found
in nature (García-Sánchez et al. 1999]). The capacity of these two minerals in terms of As adsorption will be determined
by their adsorption isotherms and the origin of the mineral (or its parent compounds
during synthesis) and the predominant species of arsenic found. Ferrous carbonate
and manganese sulfate seem to have the same ability. Other researchers already have
established a stronger association between lead and metal oxides with respect to other
ions and an adsorption of lead by the manganous oxides was up to 40 times greater
than that by the iron oxides (McKenzie 1980]). This result seems to be the case for our analysis as well; levels of As and Pb
are significantly (p??0.05) lower for ZnO (mean values of 1.3?×?10³Â ?g/kg and 55.8 mg/kg, respectively) and higher for FeO (mean values of 6.38?×?10³Â ?g/kg and 155.1 mg/kg, respectively) (Table 2). The latter mineral is of special importance as is known, in some cases, to be used
as a pigment (Potter 2000]) hence its input in compound feed may be higher relative to other minerals. Heavy
metals have been associated with a high sorption in metal oxides even in environmental
samples such as sediments (Brown and Parks 2001]).

Congruently, dicalcium- and monocalcium- phosphate samples showed the highest frequency
in contaminated samples n?=?10/22 and n?=?14/22, respectively (Table 2). In this specific case, the presence of arsenic could indicate a certain degree
of substitution between arsenate and phosphate ions in the lattice of the calcium
salt (Tawfik and Viola 2011]) or co-precipitation of the arsenate oxyanion in the presence of calcium phosphate
(Clara and Maglhães 2002]; Sahai et al. 2007]; Henke 2009]). In this case, Pb and Cd concentrations in monocalcium phosphate are significantly
higher (p??0.001 and p??0.05, respectively) than in dicalcium phosphate. The converse is true for As (p??0.001). Hg concentrations showed no significant differences between both minerals
(p??0.05).

On the other hand, despite the relatively lower concentrations of Cd found in the
samples, this metal has been reported (Chaney and Ryan 1994]; Chaney et al. 1999]; Li et al. 2005]) to have the greatest potential for transmission through the food chain at levels
that present risk to the final consumer.

One key aspect that follows from the regulatory and food safety standpoint, is the
amount of heavy metals that is ingested as a result from the mixture of several of
these mineral feeds and ingredients. For example, if a compound feed for swine is
manufactured with dicalcium phosphate and sodium chloride as metal sources exhibiting
the maximum concentration of, say, mercury detected in them and contains ca. 0.8 g/100 g
Ca and 0.4 g/ 100 g salt (a common formulation for swine nutrition and are components
with relatively high concentrations in feed) assuming this two as the only sources
of Hg then this feed will have a total of the metal of 0.38 ?g/kg. If a pig of 22 weeks
of age was fed with 4 000 g of such feed daily, then it would ingest 15 ?g mercury
every 24 h.

Even though the amount may seem small, it must be taken into account that health effects
from this substance exposure are chronic events, taking time and repeated exposure
for the contaminants to bioaccumulate up to toxic levels. Hence, animals with longer
life spans may exhibit higher concentrations of heavy metals in their tissues. As
a result of this bioaccumulation, the consumption of meat from older animals could
represent an increased risk for ingestion to a final consumer. This may suggest animals
to be vectors in heavy metal transmission along the food chain (Pagán-Rodríguez et
al. 2007]).

However, considering that several of these minerals are added to a compound feed,
in different proportions, not only the additive character of the concentrations of
these heavy metals should be addressed, but also the possibility of dilution by other
individual components with lower concentrations of said metals (e.g. maize mill) which
results in a relatively low heavy metal containing feed (Table 3). We did not find any of the samples assayed (n?=?50) [bovine (n?=?10), fish (n?=?10), poultry (n?=?10), shrimp (n?=?10) and swine (n?=?10) feeds from the most important production facilities across the country] to
surpass current legislation; in fact maximum levels of contaminants found were of
156.7 ?g/kg (Table 3).

Table 3. Heavy metal concentrations found in Costarican animal feed samples

These metals are of special importance as they are usually documented as substances
with strong toxigenic and carcinogenic capacities (van Paemel et al. 2010]). For example, both As³⁺ and As⁵⁺ are classified as group A human carcinogens (US EPA 1998]; Sapkota 2007]). This is of foremost importance in mineral feed since most vitamins and minerals
are “generally recognized as safe” according with food additive regulation (US FDA
2014]). However, some protection policies warnings to the feed industry have been issued
against the use of mineral sources that are by-products or co-products of industrial
metal production (US FDA 2003]).

Considering the data provided herein, a programme that strictly monitor the quality
on mineral ingredients and heavy metal concentrations should be implemented in order
to maintain toxic metal residues within acceptable levels and avoid contaminated feed
ingredients entering the food chain. This is especially relevant for cadmium and lead
(considering their prevalence in the environment) and mercury contemplating the relatively
elevated values found during our survey. Trends in concentration of contaminants in
feed and feed ingredients should also be observed and from the animal and human health
standpoint, interactions between heavy metals and essential nutrients (D’Souza et
al. 2003]) should be also considered.

Noteworthy, thanks to our work and the data compiled here, several Costa Rican feed
manufacturers had taken steps on improving their manufacturing practices and had avoided
all together the use of several raw materials with recurrent irregular mineral or
heavy metal concentrations. Finally, we recommend further research include speciation
as toxicity of the heavy metal involved is closely related to its oxidation state.