Processing aids in food and beverage manufacturing: Potential source of elemental and trace metal contaminants

Abstract

There are currently increased efforts to determine potential sources of trace metal contaminants in the food supply. While there are likely many sources of these elements, processing of foods has gained attention as one such route. Research is reviewed on how processing aids used for food production, including beverage filtration and fining, have been targeted as potential sources of trace metals. Potential remediation methods to reduce elemental transfer occurring during processing is discussed. While food processing aids are a critical part of food manufacturing, they can be a potential source of trace metal contaminants, including heavy metals.

Introduction

Information regarding trace metals and elemental contaminants in the food supply has increasingly been a topic of interest to both consumers and regulators. This interest is partly driven by the availability of enhanced analytical capabilities able to detect lower concentrations of trace elements in foodstuffs, along with increasing amounts of survey data on the levels of metals and trace elements in the food supply. Several governmental bodies have ongoing efforts to place defined limits on specific trace metals in foods (see Table 1). For example, a working group lead by the U.S. Food and Drug Administration (FDA) has been focusing on research involving trace elements and heavy metals in the food supply.¹ Out of these discussions, various regulations surrounding trace metals in foods have been proposed, such as a 10 μg L⁻¹ inorganic arsenic (iAs) limit for apple juice and a 100 μg kg⁻¹ iAs limit for infant rice cereal.^{1, 2} In Canada, there are also proposed regulations for As in foods, such as a limit of 15 μg L⁻¹ total As in apple juice.³ The Codex Alimentarius Commission (CAC) has published various recommendations for heavy metal limits in food and beverages, including the recommendation that lead (Pb) levels in grape juice be limited to 40 μg L⁻¹.⁴ Beyond heavy metals, there are other proposed and established limits set by the CAC and the European Food Safety Authority (EFSA) for trace metals in foods, including nickel (Ni) and chromium (Cr).^{4, 5} As such, increased information on routes of how these trace elements can enter foods and food ingredients has been a topic of increased interest to researchers.

Table 1.

Summary of key regulations and recommendation for heavy metal limits in beverages and processing aids.^a

Beverage/Processing Aid	Limit(s)	Status, Institution	Reference
Apple Juice	15 μg L−1 total As	Proposed limit, Health Canada	[3]
Apple Juice	10 μg L−1 iAs	Proposed limit, US FDA	[2]
Bentonite	2 mg kg−1 As; 1 mg kg−1 Hg; 5 mg kg−1 Pb	Current recommendation, OIV	[33]
Beverages, ready to serve	100 μg L−1 As	Current maximum level, Health Canada	[3]
Diatomaceous earth, food grade	10 mg kg−1 As; 10 mg kg−1 Pb	Current recommendation, FCC/USP	[34]
Diatomaceous earth	3 mg kg−1 As; 1 mg kg−1 Hg; 5 mg kg−1 Pb	Current recommendation, OIV	[33]
Grape juice	40 μg L−1 Pb	Current recommendation, CAC	[4]
Perlite	5 mg kg−1 As; 1 mg kg−1 Cd; 1 mg kg−1 Hg; 5 mg kg−1 Pb	Current recommendation, OIV	[33]
Water, bottled	10 μg L−1 As; 5 μg L−1 Pb	Current limit, US FDA	[21]
Wine	150 μg L−1 Pb	OIV	[33]

^aAbbreviations used: CAC, Codex Alimentarius Commission; EDTA, Ethylenediaminetetraacetic acid; FCC, Food Chemicals Codex; iAs, inorganic As; OIV, International Organization of Vine and Wine; USP, U.S. Pharmacopeia.

There are various analytical techniques available for elemental analysis of foods and beverages. Often, inductively coupled plasma is used to ionize the elements in the sample, followed by detection by mass spectrometry (ICP-MS) or optical emission spectroscopy (ICP-OES).⁶ Other analytical methods that are generally lower in cost use atomic absorption spectroscopy for detection and atomization techniques such as flame (FAAS) or hydride generation (HGAAS). While each method has particular advantages and limitations depending on the target analytes, mass spectrometry-based detection generally benefits from limits of detection that can reach as low as ng L⁻¹ concentrations. In the case of As analysis, hyphenated methods that utilize liquid chromatography are often used to separate inorganic and organic As, since iAs is associated with greater toxicity compared to the organic forms.⁷

There are likely multiple possible routes for trace elements to enter the food supply, and there are currently ongoing discussions regarding best practices to prevent such contamination from occurring.¹ Although the effect of geographical sourcing and environmental influence on trace metal levels in plant crops has been established, these factors are still not able to fully explain some of the variance observed in certain trace element concentrations in foods.^{6, 7} In turn, this has led to other aspects of the food supply chain, such as industrial and in-home processing, to be under increased scrutiny as potential sources of elemental contaminants.^8–12

For instance, processing aids are commonly used in the food industry for specific technical purposes in manufacturing. These materials are often a necessary component of food production, but they may be an overlooked source where metals can be introduced into foods. This review will evaluate the published information available on how processing aids can alter the concentrations of trace metals and elements in foodstuffs. Additionally, current recommendations on processing aids and potential remediation methods for reducing trace metal transfer to foods will be discussed.

Overview of Processing Aids

A processing aid is a substance that assists in certain technical aspects of food production in the manufacturing process. Specific examples of processing aids include antimicrobial agents used in meat processing, filter aids to process oils and beverages, and enzymes to enhance bread dough functionality. The CAC provides a four-point approach to determine what substances qualify as processing aids, with a main criterion that the substance “fulfill a certain technological purpose during treatment or processing.”¹³ FDA regulations specify a similar definition of a processing aid,¹⁴ which the U.S. Department of Agriculture Food Safety and Inspection Service (FSIS) has also adopted.¹⁵ Other governmental agencies do not have regulatory definitions for processing aids, including the United Kingdom’s (UK) Food Standards Agency.¹⁶ Processing aids are generally not required to be declared as an ingredient on a food packaging label, although use of a processing aid that contains a priority allergen or has animal origins may require proper labeling as to not inadvertently expose individuals with dietary restrictions to such substances.¹⁷ Since it is not always practical or possible to remove the processing aid after it accomplishes its intended purpose, the CAC acknowledges that use of a processing aid can unintentionally introduce a “residue” into the final processed material.¹³ Because of this possible unintentional introduction of chemical components, it is important to understand the downstream effects of using a processing aid and to determine the concentrations of any substances remaining in the food from processing.

Effect of Processing Aids on the Occurrence of Trace Metals in Foods and Beverages

Past research demonstrates that various processing aids can have a significant impact on the elemental composition of food and beverages (see Table 2). Filter aids such as diatomaceous earth (DE) and perlite are a type of processing aid and are often used in beverage processing to increase throughput by preventing a filter element from prematurely fouling from suspended solids present in the liquid phase.¹⁸ Additionally, bentonite clay is often used for beverage clarification purposes.¹⁹ Since DE, perlite, and bentonite are derived from mined materials, they can contain a large array of elements, including heavy metals. Other processing aids, such as alumina, can be used to process water to reduce levels of certain target elements.²⁰

Table 2.

Summary of research on the effect of processing aids on the occurrence of metals and trace elements in food and beverages.^a

Processing aid	Processed food/beverage	Effect of processing on elemental concentrations	Analytical method	Author, year
Alumina	Water	Increase in Al, B, Br	--	EFSA, 2006
Bentonite	White wine	Increase in Al, Be, Ce, Zn, Ga, Y, Zr, Nb, La, Pr, Nd, Dy, Ho, Yb, Hf, Th, and U	ICP-MS	Gomez et al., 2004
Bentonite	White wine	Increase in Al, Ba, Be, Ca, Ce, Co, Cs, Fe, Ga, Gd, La, Li, Mg, Mn, Na, Nd, Ni, Pr, Sb, Sr, Tl, U, V, Y; decrease in Cu, K, Rb, Zn; No change in Pb	ICP-MS/ICP-OES	Nicolini et al., 2004
Bentonite	Wine	Increase in Al, As, Ba, Be, Bi, Ca, Co, Cd, Fe, Ga, Ge, Li, Mg, Mn, Mo, Na, Nb, Ni, Sb, Sc, Sn, Sr, Tl, V, W, Y, and Zr; decrease in B, Cu, K, Rb, and Zn	Na, Mg, K, Ca, and Fe was determined by FAAS; Al by ET-AAS; all other analytes by ICP-MS	Catarino et al., 2008
Bentonite	Grape syrup (pekmez)	Increase in Cd and Pb; decrease in As, Cu, Fe, Zn	GFAAS	Heshmati et al., 2020
Diatomaceous earth	Sake	Increase in As and Pb	--	Namba et al., 1981
Diatomaceous earth	Maple syrup	Decrease in Pb	GFAAS	Stilwell and Musante, 1996
Diatomaceous earth/cellulose blend	Red table wine	Increase in Pb	ICP-MS	Almeida and Vasconcelos, 2003
Diatomaceous earth	Beer	Increase in Al, As, Ca, Fe	As was determined using HG-AAS; all other analytes by ICP-OES	Coelhan, 2014
Diatomaceous earth	Brandy	Increase in Ca and Si	ICP-OES	Gomez et al., 2014 and 2015
Diatomaceous earth	Apple and grape juice	Increase in iAs and Pb	ICP-MS; LC-ICP-MS for As speciation	Wang et al., 2017
Diatomaceous earth/perlite blend	Apple juice	Increase in V	ICP-MS	May et al., 2019
Diatomaceous earth/activated carbon blend	Beeswax	Increase in total As and Si; decrease in Ca, Fe, Hg, Mn, P, Zn	ICP-MS	Navarro-Hortal et al., 2019
Diatomaceous earth	Beer and wine	Increase in iAs and Pb	ICP-MS; LC-ICP-MS for As speciation	Redan et al., 2019
Perlite	Sake	Increase in As	--	Namba et al., 1981
Perlite	Apple and grape juice	No change in iAs; decrease in Pb	ICP-MS; LC-ICP-MS for As speciation	Wang et al., 2017

^aAbbreviations used: iAs, inorganic As; ICP-MS, inductively coupled plasma-mass spectrometry; LC-ICP-MS, liquid chromatography-inductively coupled plasma-mass spectrometry; HGAAS, hydride generation-atomic absorption spectroscopy; ICP-OES, inductively coupled plasma-optical omission spectroscopy; ETAAS, electrothermal atomic absorption spectroscopy; FAAS, flame atomic absorption spectroscopy; GFAAS, graphite furnace atomic absorption spectroscopy

Alumina

Due to interest in processing aids potentially releasing trace metals, a scientific panel convened by the European Food Safety Authority (EFSA) addressed concerns on this topic.²⁰ Oxyhydroxide-containing processing aids can be used as a filtration media to remove elevated fluoride ions from spring water, which may be needed to keep levels from exceeding bottled water regulatory limits for flouride.²¹ Due to the composition of the filter media, there was a question on whether trace metal release can occur from the media. The EFSA panel determined that, based on data submitted to the panel, metal concentrations in water increased from 18 to 86 μg L⁻¹ for aluminum (Al), 160 to 280 μg L⁻¹ for boron (B), and 66 to 550 μg L⁻¹ for bromine (Br) after processing with the oxyhydroxide media. These data indicated that use of a processing aid typically considered to be inert is able to alter concentrations of various trace elements in water.

Bentonite

Research on how processing affects trace metal concentrations in beverages arose from studies seeking to determine how heavy metals can transfer to wine throughout its production. In particular, use of bentonite clay as a processing aid has been the subject of various studies since it is commonly used as a fining agent by the wine industry to promote beverage clarity. Bentonite can contain elevated levels of metals and trace elements that can subsequently transfer to the processed product.⁶ Catarino et al. tested how use of bentonite in wine processing can affect the levels of various target elements in Portuguese wine.²² For their experiments, the authors mixed 5 g of bentonite with 200 mL wine and then left the slurry to settle for 24 h in order to simulate the fining process. Analysis of the supernatant (clarified wine) showed significant increases in 28 different elements, including the heavy metals As, Pb, and Cd. Most notably, As levels in wine increased more than 30-fold (up to 38 μg L⁻¹) when bentonite containing an As concentration of 38 mg kg⁻¹ was used. The quantity of bentonite utilized in this study for fining was likely greater than that would be used in a commercial setting, but the results still demonstrate the extent of trace metal release from this processing aid.

In another study aimed at determining how processing with bentonite affects trace metal levels in wine, Gomez et al. tracked 63 elements throughout the winemaking process in German white table wine.²³ The authors performed elemental analysis of the wine prepared under two different fermentation methods, where the bentonite was added in either before or after the fermentation process. When bentonite was added to wine after fermentation for clarification, there were increases in both As and Pb levels. In contrast, As and Pb concentrations in wine were lower when bentonite added to the grape must prior to fermentation. The authors hypothesized that the observed decrease in heavy metals may have been due to the formation of insoluble metal complexes that then precipitated out of solution.

Similarly, Nicolini et al. conducted experiments to assess how processing wine with different types of bentonite influences the final elemental composition of white wine.¹⁹ The authors tested this by hydrating 45 mg of select types of bentonite with 450 μL water, and then combining this slurry with 45 mL white wine. The slurry was mixed for 3 h, centrifuged, and then the supernatant was removed for analysis. The results showed that there was a significant effect of bentonite type on wine Pb concentrations, with levels increasing as much as 70% to a concentration of 54.8 μg L⁻¹. These results are consistent with previous studies indicating that use of bentonite as a processing aid may affect elemental composition in wine.

Heshmati et al. (2020) tested how processing grape syrup (pekmez) with bentonite affected levels of target metals in the product.²⁴ In this study, the authors clarified the grape juice used to produce pekmez by the addition of bentonite (1.5–3% by weight) containing 40 mg kg⁻¹ Pb, 0.04 mg kg⁻¹, and 0.1 mg kg⁻¹ As. The sample was agitated the sample for 1 h and then left to settle for 24 h before collecting the supernatant, drying the sample, and then performing analysis for target elements by GFAAS. The results found that use of bentonite increased Cd and Pb levels in the dried sample by ~50–100% up to ~0.07 μg g⁻¹ and ~ 0.031 μg g⁻¹ for Pb and Cd, respectively. In contrast, levels of As, Cu, Fe, and Zn decreased by 20–80%.

Diatomaceous Earth and Perlite

One of the early studies on how processing aids can affect the levels of metals in foodstuffs tested how filtration aids can alter trace metals in beverages. In this report, Nambe et al. filtered sake (a rice wine) by using various types of diatomaceous earth (DE) or perlite filter aids and then analyzed the processed sake for elements, including Pb and As.²⁵ The sake filtered with DE contained Pb levels ranging from 0.1–2.6 mg L⁻¹ and As from 1–5 mg L⁻¹. Sake filtered with perlite contained 4–5 mg L⁻¹ As after filtration; in contrast, there was no detectable Pb in the sake processed with perlite. Increases in iron (Fe) levels from use of all filter aids were observed, which is of concern because it can affect product quality by promoting oxidation of important flavor compounds. These results were an early indicator that the type of filter aid used in processing can have an impact on the final trace metal concentration in sake and possibly other beverages.

In another study from Portugal, Almeida and Vasconcelos conducted a study that tracked Pb throughout the winemaking process. The researchers performed Pb analysis of red wine along the major vinification steps, including after initial pressing, fermentation, and filtration in order to help identify any major contributors to Pb levels in wine. The experiments found an increase, although small, in Pb concentrations in red table wine after filtration using a DE/cellulose blend as a filter aid.²⁶ Although the elemental composition of the filter aid was not reported, this study suggests that filtration treatments can affect Pb levels in wine.

A series of two publications by Gomez et al. found that filtration of distilled spirits using DE resulted in increased silicon (Si) and calcium (Ca) concentrations.^{27, 28} Although not known to be associated with toxicity-related endpoints, excessive Si levels can drive precipitate formation in the beverage, leading to an undesirable aesthetic appearance. The study authors first performed a laboratory-scale experiment by mixing 10 g DE with 200 mL of a brandy simulant (36% etOH by volume) for 30 min, separating the solids by centrifugation, and then analyzing the supernatant by inductively coupled plasma-optical omission spectroscopy (ICP-OES). Subsequently, an industrial-scale filtration trial was performed where Brandy de Jerez was processed using a similar ratio of filter aid as in the laboratory-scale experiments. Although the authors mention that DE is generally considered to be an inert material, the results indicated that Si levels in the brandy simulant contained 480 μg kg⁻¹ after being processed in the laboratory-scale experiment. Brandy that was filtered using industrial equipment increased from 180 μg kg⁻¹ (unfiltered control) up to 240 μg kg⁻¹ Si. Increases in Ca levels were also observed, although they did not exceed 1 mg kg⁻¹ under both laboratory and industrial filtration conditions.

In addition to wine and distilled spirits, there has been continuing research on how processing aids can influence trace metal levels in other beverages. For example, one report from Germany drew attention to the issue of filtration aids elevating heavy metal concentrations in beer.²⁹ Even though there are strict centuries-old laws regulating how beer is produced in Germany, periodic analysis of German beers resulted in detection of elevated As levels. Since it was unlikely that any of the raw ingredients involved in beer brewing would contribute to elevated heavy metal concentrations, it was hypothesized that processing aids used for beer filtration were a potential source of these metals. To test this hypothesis, 2% slurries of different types of DE in beer were mixed for 5 min and then separated to analyze the liquid fraction for any potential metal release from the DE.²⁹ Trace metals analysis of the beer indicated that direct mixing of the beer with DE resulted in increases of As, Fe, Al, and Ca content of the beer. These experiments indicated that use of filter aids for beer processing was able to alter its elemental concentration, including heavy metals.

Further research on how filter aids can result in alterations of trace metals in beverages was conducted by Wang et al. using a laboratory-scale filter to simulate filtration treatments that occur in an industrial setting.¹⁸ In these experiments, unfiltered apple and grape juice were processed using a leaf-filter with DE or perlite as filter aids. The amount of filter aid used in processing was typically less than 1 g L⁻¹ and was based on the suspended solids content of the beverage. Analysis of the processed juice using inductively coupled plasma-mass spectrometry (ICP-MS) for target heavy metals (As, Pb, and Cd) found that use of DE containing 3–7 mg kg⁻¹ iAs increased levels of total As from 4.7 up to ~27 μg L⁻¹ in apple juice and from 6.1 to ~10 μg L⁻¹ in grape juice. Juices processed with perlite containing less than 1 mg kg⁻¹ iAs did not result in a significant increase in total As concentrations. Further analysis of the processed juice using liquid chromatography (LC) in line with ICP-MS found that the observed increase in As levels were driven by increases in the inorganic forms (As^III and As^V) of the heavy metal. Results from the Pb and Cd analyses were more difficult to interpret due to high variance in the data, which may have been due to heterogeneous distribution of heavy metals throughout the lots of filter aids used in the experiments. Still, this study demonstrated that use of filter aids altered levels of target heavy metals in juices when processed using a laboratory-scale leaf filter.

Research by Redan et al. then tested whether processing unfiltered beer and wine with DE using a laboratory-scale leaf filter produced results that were similar to those observed with processing apple and grape juices.⁷ For these experiments, two types of beer (ale- and larger-style) and wine (red and white) underwent filtration treatments with three different types of DE. The experimental design of the study assessed whether trace metal transfer was affected by the quantity of filter aid and the filtration time. Beer and wine were processed with DE amounts ranging from 1–5 g L⁻¹, and then were analyzed by ICP-MS for total As, Pb, and Cd. LC-ICP-MS was used for As speciation. The results found that there was a direct relationship between the quantity of DE used in filtration and the amount of As transferred to the filtered beverage, with levels increasing by 11–14 μg L⁻¹ iAs in the processed beverages when 2 g L⁻¹ DE was used. Pb and Cd were also monitored, but increases in Pb were only detected when higher amounts (10 g L⁻¹ DE) of the filtration aid were used, and Cd levels did not significantly increase. Minimal transfer of Pb and Cd to beer and wines are likely due to the relatively low levels of soluble Pb (ranging from 0.1–0.2 mg kg⁻¹) and Cd (ranging from not detectable to 0.03 mg kg⁻¹) in the filter aids. An assessment of the transfer efficiency of iAs and Pb from DE to beer and wine indicated that both were highly soluble, as >80% of total iAs and >70% of total Pb transferred from the DE after being mixed with beer or wine for 2 h. The results from this study on beer and wine filtration are consistent with the work of Wang et al. on juices, indicating that use of filter aids in processing can affect heavy metal concentrations across various beverage types.

Further, in a study performed in Germany, May et al. tested the transfer of trace metals from DE to juice when processed in an industrial setting.³⁰ The authors first performed elemental analysis of both clarified and unfiltered commercial apple juice and found that the clarified juice had significantly higher vanadium (V) concentrations when compared to the unfiltered juice. In order to determine whether processing was responsible for the difference in V levels, the authors performed a series of experiments where they processed unfiltered apple juice using a filter aid containing varying V levels. For the experiments, apple juice was processed using a candle filter with 1.5 g L⁻¹ DE containing V concentrations ranging from 38–370 mg kg⁻¹. The authors then analyzed the juices using ICP-MS and found that apple juice V concentrations increased from ~3 μg kg⁻¹ in the unfiltered juice to as much as 200 μg kg⁻¹ after filtration. The data showed that the variance in the filtered juice V concentrations were predicted (R² > 0.99) by the amount of V in the filter aid blend. Although the authors concluded that the V levels in the filtered juice did not pose a concern from toxicity-related endpoints, these results are still important when considering that use of DE in processing can be a major contributor of certain trace metals in juices.

In addition to beverages, filter aids are commonly used to process other foodstuffs. Navarro-Hortal et al. processed beeswax using an DE/activated carbon blend and found that it altered the levels of trace metals in the filtered material.³¹ The authors mixed beeswax with filter aid at a ratio of 1 g per kg beeswax and then filtered the slurry by pumping it through a cellulose filter. The authors found that such processing significantly altered the levels of multiple target elements in the final product. While As levels increased from 15 to 48 μg kg⁻¹, Hg concentrations decreased from 1.7 to 0.1 mg kg⁻¹. Cd and Pb levels were not significantly affected by filtration. These results indicate that processing with filter aids has potential to either increase or reduce levels of heavy metals in some foods.

An additional report demonstrated that use of DE can reduce heavy metals levels in food products. In this study, Stilwell and Musante performed a study assessing Pb transfer to maple syrup throughout its production.³² Samples for analysis were selected at points directly after the maple sap was collected at the maple grove until after the sap was filtered and placed in a stainless steel holding tank. Sample analysis indicated that Pb concentrations in sap decreased from approximately 30 to 20 μg L⁻¹ after filtration using DE. The authors postulated that since Pb can form an insoluble precipitate in the presence of phosphates naturally occurring in maple sap, filtration was able to reduce Pb levels by removing precipitate that may have formed.

Methods to Limit Trace Metal Transfer from Processing Aids in Food and Beverage Manufacturing

Several studies have reported specific steps that are able to limit metal release from processing aids (see Table 3). For example, Namba et al. found that washing DE, perlite, and cellulose filter aids with a 9.7% HCl solution was able to reduce Pb and As levels in these materials.²⁵ Besides mineral acids, organic acid solutions (for example, citric and malic acids) have also been shown to reduce heavy metal levels in filter aids. In experiments performed by Wang et al., DE containing ~7 mg kg⁻¹ iAs was washed by preparing a 10% slurry composed of DE/water and then adjusting the pH of the slurry from 2.6–3.0 with citric acid, malic acid, or the chelating agent ethylenediaminetetraacetic acid (EDTA). After mixing the slurry for 25 min, the DE was collected and then analyzed for heavy metals. The results found that washing with solutions of citric acid, malic acid, EDTA, or even water alone, all reduced iAs levels in DE to less than 2 mg kg⁻¹. However, Pb or Cd levels in DE were not affected. Further experiments found that use of DE washed with citric acid to process apple and grape juices did not increase iAs levels in the filtered juices.

Table 3.

Summary of research on potential remediation wash methods to reduce metals in processing aids.^a

Processing Aid	Wash solution	Effect of washing on processing aid	Author, year
Diatomaceous earth	9.7% HCl	Decrease in Fe and Pb	Namba et al., 1981
Diatomaceous earth	citric acid, malic acid, EDTA, or water	Decrease in iAs	Wang et al., 2017
Diatomaceous earth	0.1% citric acid, 0.1% EDTA, or water	All wash solutions decreased iAs; EDTA solution decreased Pb	Redan et al., 2019
Perlite	9.7% HCl	Decrease in Fe and Pb	Namba et al., 1981

^aAbbreviations used: EDTA, Ethylenediaminetetraacetic acid; iAs, inorganic As.

A similar experiment was performed by Redan et al. where DE containing ~6 mg kg⁻¹ iAs was washed by mixing a 10% slurry of filter aid with either a 0.1% solution of citric acid, EDTA, or water alone for 1 min.⁷ Analysis of the DE indicated that washing using the citric acid solution, EDTA solution, or water was able to reduce levels of iAs in DE to ~2 mg kg⁻¹ or lower. Pb concentrations in DE (0.12 mg kg⁻¹) were reduced by 50% after being washed with the EDTA solution, while Cd in DE was not affected by any treatment.

There have been some efforts to identify specific beverage properties that influence trace metal transfer from filter aids. In one study, model juices with pH values ranging from 3–4 and °Brix values from ~8–18 were processed using DE as a filter aid. The results found no significant effect of either pH or °Brix on iAs release from DE. Another study filtered beer and wine with pH values ranging from ~3–5 and found a modest effect of pH on iAs transfer.⁷ In addition, an EFSA report on processing aids reported an effect of pH on Al release from an oxyhydroxide filter media.²⁰ The results showed that Al release from the filter media was dependent on pH, with certain conditions increasing Al concentrations in the filtered water in the range of 100–200 μg L⁻¹, while pH control (exact pH levels were not indicated) limited Al levels to 40–60 μg L⁻¹. Together, these results indicate that the chemical properties of beverages, including pH, may have an impact on metal release from processing aids.

Current Recommendations on use of Processing Aids in Foods and Beverages

As indicated by the CAC, use of a processing aid in food manufacturing may unintentionally introduce substances into foodstuffs even after the processing aid is removed. Hence, the CAC gives the recommendation that processing aids be of food-grade quality as determined by a reputable organization. An example of an organization setting recommendations for processing aids includes the International Organization of Vine and Wine (OIV), with limits for soluble metals in bentonite set at 5, 2, and 1 mg kg⁻¹ for Pb, As, and Hg, respectively.^{22, 33} The Food Chemicals Codex (FCC) compendium includes a specification for food-grade DE and has the recommendation that it contain less than 10 mg kg⁻¹ soluble As and Pb when used as a filter aid for food and beverages.³⁴ It is important to note that the studies performed by Redan et al. and Wang et al. for filtration of alcoholic beverages and juices both used food-grade quality filter aids in their experiments. Under certain conditions, As concentrations in the filtered beverages exceeded the limit proposed by the FDA for apple juice (10 μg L⁻¹ iAs) or the total As limit proposed by Health Canada for apple juice (15 μg L⁻¹). These data may indicate that the current recommendations by USP for food-grade DE need to be re-evaluated to reflect changes in regulations for heavy metals in foods.

Summary and Future Research

Use of processing aids is necessary for many types of food processing techniques, including filtration of foods and beverages. A review of the published literature found that there can be elemental transfer from processing aids to foodstuffs and indicates that these materials may be an important factor to consider when addressing limits for trace metals in foods. In particular, use of filter aids and fining agents that are typically composed of mined materials (e.g., DE and bentonite) have been reported to release trace elements, including heavy metals. Much of previous research is limited by a lack of analysis on the actual content of the soluble trace metals in the processing aid used in the experiments. Because of this, it is often difficult to directly ascertain the exact contribution of trace elements to foodstuffs from the processing aid material itself. Further research should be conducted on addressing the solubility of trace metals and elements from these processing aids in order to more fully characterize the degree to which the trace metals in processing aids is present in the final product available to the consumer. Together, this information can help manufacturers, governmental bodies, and organizations that publish compendial standards to develop best practices to limit dietary exposure to potentially toxic metals.

Abbreviations Used

CAC

Codex Alimentarius Commission

DE

diatomaceous earth

EFSA

European Food Safety Authority

EDTA

ethylenediaminetetraacetic acid

FCC

Food Chemicals Codex

ICP-MS

inductively coupled plasma-mass spectrometry

ICP-OES

inductively coupled plasma-optical omission spectroscopy

OIV

International Organization of Vine and Wine

UK

United Kingdom

USDA-FSIS

U.S. Department of Agriculture Food Safety and Inspection Service

US FDA

U.S. Food and Drug Administration

USP

U.S. Pharmacopeia