Food Additives & Contaminants: Part A

ISSN: 1944-0049 (Print) 1944-0057 (Online) Journal homepage: http://www.tandfonline.com/loi/tfac20

Mineral oil hydrocarbons in food - a review

Koni Grob

To cite this article: Koni Grob (2018): Mineral oil hydrocarbons in food - a review, Food Additives
& Contaminants: Part A, DOI: 10.1080/19440049.2018.1488185

To link to this article: https://doi.org/10.1080/19440049.2018.1488185

Accepted author version posted online: 12

Jun 2018.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=tfac20
Mineral oil hydrocarbons in food - a review

Koni Grob*

Kantonales Labor Zürich (Official Food Control Authority of the Canton of Zürich), P.O. Box,
CH-8032 Zürich, Switzerland

t
ip
Abstract

cr
Work on mineral oil hydrocarbons (MOH) contaminating food is reviewed up to about 2010,
when the subject received broad publicity. It covers the period of the main discoveries and

us
elimination or reduction of the dominant sources: release agents used in industrial bakeries,
spraying of rice, additions to animal feed, contamination of edible oils from various sources
and migration from paperboard packaging. In most cases highly refined (“white”) oils were
an
involved, but also technical oils, e.g. from the environment, and more or less crude oil
fractions from jute and sisal bags. There were numerous unexpected sources, and there
M
might still be more of those. The exposure of the consumers to MOH must have been
markedly reduced in the meantime. Environmental influx may have become dominant,
particularly when taking into account that these MOH go through several degradation
ed

processes which might enrich the species resisting metabolic elimination. Major gaps are in
the systematic investigation of sources and the largely unavoidable levels from
environmental contamination, but also in the toxicological evaluation of the various types of
pt

hydrocarbons. A regulation is overdue that avoids the present discrepancy between the low
tolerance to MOH perceived as contaminants and the very high legal limits for some
ce

applications – the MOH are largely the same.

Ac

Keywords: Mineral oil saturated hydrocarbons (MOSH); Mineral oil aromatic hydrocarbons
(MOAH); Release agents; Batching oil; Printing inks; Environmental pollution; Animal feed

Introduction
Food contamination with mineral oil hydrocarbons (MOH) became a subject of interest in
the early nineteen nineties. At that time, “white” or “well purified” mineral oils were often
called “food grade”, suggesting that they were quasi edible. “White” oils are largely free of

CONTACT Koni Grob E-mail: [email protected]. Kantonales Labor Zürich (Official Food Control Authority
of the Canton of Zürich), P.O. Box, CH-8032 Zurich, Switzerland.

1
mineral oil aromatic hydrocarbons (MOAH), hence mainly consist of mineral oil saturated
hydrocarbons (MOSH). There was, however, little justification for this – on the contrary,
frequent granuloma formation in human tissues was ascribed to mineral oil as from the
1960s (e.g. Dincsoy et al., 1982; Cruickshank, 1984). In fact, in 1989, the European Scientific
Committee on Food (SCF) stated that ‘‘there is no toxicological justification for the continued
use of mineral hydrocarbons as food additives’’ and specified temporary acceptable daily
intakes (ADIs) of 0–0.005 mg/kg and 0–0.05 mg/kg body weight (bw) for oleum-treated and
hydrogenated mineral oils, respectively (SCF, 1989). However, the limits that would normally

t
have been derived (0.3–3 mg/kg food) were not imposed; as shown below, concentrations

ip
often exceeded these by 100-1000 times.

Animal testing, largely performed by the mineral oil industry during the nineteen nineties,

cr
produced puzzling results, but a low toxicity of saturated hydrocarbons of high molecular
mass was derived that resulted in far more tolerant evaluations by the SCF (SCF, 1995) and

us
the Joint FAO/WHO Expert Committee on Food Additives (JECFA, 2002). However, this
evaluation was soon questioned by Scotter et al. (2003) and Noti et al. (2003), since the
an
hydrocarbons most strongly accumulated in human tissue just fall into the mass range with
the high ADIs. This contradiction persists today.

To avoid accumulation of MOSH, the hydrocarbons should either be below C20 (probably
M
exhaled) or above around C40 (to prevent absorption) – which is just not the typically used
oil considered as food grade. Accumulation in tissues was the base for the evaluation by the
ed

German Federal Institute for Risk Assessment (BfR), setting a maximum value of 12 mg/kg
food for MOSH in the range of C10 to C16 (BfR, 2015) and of 4 mg/kg food for the range of
C17 to C20 (BfR, 2012). Consequently, the limits should even be lower for the MOSH above
pt

C20, but this would have been conflicting with the ADIs of the SCF and JECFA.

In an extensive evaluation by the CONTAM Panel of EFSA it was concluded that the
ce

present exposure to MOSH (estimated as 0.03-0.3 mg/kg bw/d) is of “potential concern”

(EFSA, 2012). This is orders of magnitude below the exposure to be expected if the ADIs
(10-20 mg/kg bw/d) were exploited, but this conflict was not highlighted.
Ac

In 2013, for food additives, the ANS Panel of EFSA surprisingly returned to the high ADI
specified by JECFA, considering a potential dietary intake of 10.1 mg/kg bw/day by toddlers
acceptable, even though it exceeds the present exposure for adults estimated by the EFSA
CONTAM Panel 34-340 times. Later data confirmed that a far reaching revision of the
regulations is needed (Barp et al., 2014; Barp et al., 2017a,b). Independently of these
conflicting scientific evaluations, the public perception changed from a “food-grade”
component to a contaminant of concern.

2
 The history of mineral oil analysis in food is strongly influence by the analytical technology
that promoted, but in a way also hindered progress. On-line coupled HPLC-GC-FID is the
method of choice. It was largely developed in Kantonales Labor Zürich (KLZ), making use of
solvent effects and retention gap technology. The first version was successful in 1984 (Grob
et al., 1984), the first automated instrumentation becoming available on the market in 1990.
Raw food extracts or solutions of edible oils could be directly injected, which not only
provided pre-separation at high performance, but also avoided contamination during sample
preparation with virtually ubiquitous MOH. However, up to around 2010, this technique found

t
its way into hardly any control laboratory and, therefore, the KLZ remained alone in analyzing

ip
mineral hydrocarbons in food, with correspondingly limited impact.

Up to 2008, the on-line HPLC-GC method was limited to the analysis of the saturated

cr
hydrocarbons (Grob et al., 1991a). The composition of the mineral oil product detected
(content of aromatics, heteroatom compounds and additives) was derived from the source.

us
Only after the crisis of the Ukrainian sunflower oil, in which the presence of aromatics was
neglected (Biedermann and Grob, 2009), the method was upgraded to enable routine
an
measurement of the aromatic hydrocarbons (Biedermann et al., 2009). Hence, the data
available up to 2008 were restricted to MOSH (the terms MOSH and MOAH were proposed
in that paper).
M
During most of the period, the detection limit was in the range of several mg/kg food, then
considered sufficient, since concentrations found were often more than two orders of
ed

magnitude higher. In the meantime, such high contamination virtually disappeared and the
detection limits were lowered, if necessary by enrichment or removal of interfering
components (Zurfluh et al., 2014).
pt

Heimbach et al. (2002) estimated the dietary exposure of US consumers to MOH by

multiplying (then scarce) MOH concentration data with amounts of foods consumed,
ce

assuming that all foods of a given type were contaminated at that level. For the sum of all
types of MOH, they came up with 0.875 mg/kg bw/d. Of this, 0.427 mg/kg bw/d were white
oils used as pan-release agents in baking, for de-dusting of stored grain, in confectioneries,
Ac

and in coatings for fruits and vegetables; 0.404 mg/kg bw/d consisted of petrolatum, primarily
used as trough grease in bakery applications. Exposure to waxes was 0.044 mg/kg bw/d,
mainly from fruit and vegetable coating, followed by flexible packaging.

For European consumers, Tennant (2004) estimated the daily dietary exposure to mineral
oils as 0.39-0.91 mg/kg bw/d for adults and 0.75-1.77 mg/kg bw/d for children (mean and
97.5th percentiles). Wax intake was estimated as 0.08-0.19 mg/kg bw/day for adults and
0.23-0.64 mg/kg bw/d for children. When usage factors were applied, estimates of chronic
intakes of white oils were reduced to 0.09-0.20 mg/kg bw/d for adults and to 0.17-0.39 mg/kg
bw/d for children, those of waxes to 0.01-0.02 mg/kg bw/d for adults and 0.02-0.06 mg/kg

3
bw/d for children. For white mineral oils, the principal source was assumed to be imported
de-dusted grain; for waxes it was glazing agents on confectionery. The database for these
estimates was rather scarce (and not supported by the data reported below), but the values
were in the same range as those estimated by EFSA (EFSA, 2012).

This paper summarizes findings up to about 2010, when MOH became a public issue, and
covers the perhaps less known period of the discovery of the main sources. It reviews
published data from those times, but also includes unpublished data from the KLZ, i.e. from
the Swiss market. It should support the often tedious search for the source of a

t
contamination, but also provide some indications on the decrease of the exposure since the

ip
nineteen nineties. The sequence in which the findings are presented follows the time when a
given subject was in the focus of the investigations.

cr
Jute and sisal bags

us
Hazelnuts

In the KLZ, the investigation of MOH in foods started from an incidental finding. In 1988,
an
there was much interest in irradiating foods and the determination of such treatment due to
required labeling. The alkenes cleaved from lipids were analyzed by on-line HPLC-GC
M
(Biedermann et al., 1989). For some reasons, also hazelnuts were analyzed and found to
contain hydrocarbons typical for mineral oil, including pristane and phytane. All hazelnuts
from the Swiss market contained MOH, but not those from domestic gardens. After
ed

protracted search, the source turned out to be the jute bags in which they were imported
(mainly from Turkey, some from Italy).

Jute (as well as sisal) bags are made of fibers that must be treated with a batching oil
pt

(then normally a brownish, low grade mineral oil) to improve spinning. Related to the fibers,
5-7 % batching oil is applied, of which roughly half evaporates during subsequent airing (to
ce

avoid an off-flavor of the packed food). Typically bags of approximately 1.5 kg are used for
packing 50 kg food, i.e. approximately 45 g oil are in contact with the food (hypothetical total
Ac

transfer would result in concentrations around 900 mg/kg food). Roughly a third of the MOH
is of a volatility to be transferred to dry foods. Concentrations in hazelnuts mostly ranged
between 10 and 50 mg/kg; the maximum was 500 mg/kg, possibly from nuts located close to
the wall of the bag (Grob et al., 1991).

The problem was solved rather rapidly: since there was a direct contact between the
importers and the packers, it was written into the contracts that mineral-oil-free jute bags
must be used, as were already available from India. In 1991, a legal limit of 10 mg/kg for all
types of nuts was introduced into Swiss legislation.

4
Rice

In 1990, a substantial part of the rice, particularly that from Asia, was imported in jute
bags. While European rice, imported in bulk mainly from Italy, commonly contained less than
2 mg/kg MOH (1.1-1.6 mg/kg in 6 samples analyzed in 2002), the mean value for Asian rice
was around 100 mg/kg (6 samples), with a maximum of 160 mg/kg (Grob et al., 1991a).

Fig. 1 compares the extract from the jute bag with that from the rice that was packed in
the bag. While the saturated hydrocarbons in the jute bag ranged from n-C14 to n-C35, those
in the rice only reached up to n-C23, as typical for transfer through the gas phase (Lorenzini

t
et al., 2010). Of the most volatile hydrocarbons, the transfer was almost complete. Rice also

ip
contains natural hydrocarbons, mainly odd-numbered n-alkanes, which are easily

cr
distinguished from the mineral n-alkanes with an equal distribution of even- and odd-
numbered species. The composition of the batching oils varied in terms of width of the mass

us
distribution and abundance of the n-alkanes (n-alkanes largely being degraded or digested in
some bags).

In 1993, after some measures had been taken, the average batching oil concentration in
an
35 samples of Asian rice from the Swiss market was 26 mg/kg, with a maximum of 100
mg/kg for a Basmati rice (three others of 42, 45 and 80 mg/kg).
M

Fig. 1. On-line HPLC-GC-FID of extracts from a jute bag and the rice that had been
packed in the bag (adapted from Grob et al., 1991a).
ed

Even today, Asian rice is rather frequently contaminated with batching oils, but
pt

concentrations are seldom above 10 mg/kg (Biedermann and Grob, 2015). This is due to
farmers sending the rice to the trader in jute bags made with mineral batching oil.
ce

Contamination is lower, since the bags are re-used. According to a manufacturer, the
majority of the jute bags are still made with mineral batching oil, though no longer those sent
to Europe.
Ac

Chocolate

In 1990, cocoa beans came to Europe in jute and sisal bags and, indeed, chocolate
regularly contained 10-50 mg/kg batching oil hydrocarbons. Considering that cocoa butter
and cocoa powder constitute roughly a third of chocolate, MOH concentrations in the added
cocoa components must have been roughly three times higher.

In 1990, 110 and 120 mg/kg batching oil were measured in the pseudo-shells of two lots
of roasted beans and 8 and 13 mg/kg, respectively, in the kernels, indicating protection of the
kernel. However, this concentration was just about to increase, since the processing was

5
adjusted to reduce microbial contamination: before roasting, beans were steam-treated,
which strongly promoted the migration of the MOH to the kernels. Of 10 chocolate bars, two
contained 160 and 270 mg/kg batching oil (the others 5-15 mg/kg; Grob et al., 1991), which
means that the cocoa part must have contained more than 800 mg/kg batching oil. In 1993,
the mean MOH concentration in 30 samples of chocolate was 22 mg/kg, with a maximum of
75 mg/kg; in 1994 the values were 24 mg/kg and 77 mg/kg, respectively (35 samples).

For recycled paperboard, it was determined that migration through the gas phase into dry
foods was substantial up to hydrocarbons eluted from GC at the retention time of n-C24

t
(Lorenzini et al, 2010), but in chocolate and other cocoa products this upper limit sometimes

ip
exceeded n-C30 (e.g. Grob et al., 1991a), presumably because of the high temperatures in
the countries of origin.

cr
Coffee

us
Around 1990, many more foods were imported in jute or sisal bags than today. Roasted
and ground coffee beans commonly contained around 100 mg/kg mineral batching oil, with a
an
maximum of 240 mg/kg (Grob et al., 1991). From the molecular mass distribution, in
particular the most volatile constituents, it was concluded that the roasting process removed
little, but no MOH were found in brewed coffee at a detection limit of about 10 mg/L. Instant
M
coffee only contained up to 2 mg/kg mineral oil, presumably for the same reason: extraction
with water left the hydrocarbons in the residue.
ed

Better batching oils

Although batching oils based on plant lipids were available (e.g. used for bags to pack
hazelnuts), solving the problem for chocolate and other commodities was lengthy because of
pt

the trade through spot markets and lack of direct contacts to producers. The available
measures were ineffective against international trade agreements (the KLZ is only competent
ce

for the local market). In addition, the major producers of jute bags were in Bangladesh, at
that time known to be an extremely poor country with jute being the main export article.
Ac

Malnutrition definitely being a health risk, the KLZ avoided harsh measures. Furthermore,
around 1990, the jute bags were popular (“jute instead of plastic” being a common slogan);
the distinct odor was attributed to natural constituents of the fibers. The message that they
contaminate food was correspondingly difficult to communicate.

Some improvement would have been possible by reusing the bags. However, properly
opening the bags and sending them back was more expensive than making new ones plus
the costs for disposal (a bag produced in Bangladesh was said to cost €1.20; €1.60 with a
vegetable batching oil).

6
 In 1991, in a first attempt to increase pressure, a sample of batching oil was analyzed for
the aromatics. Using on-line HPLC-GC-FID in backflush for the aromatics, the sum
amounted to 23 % of the oil. Furthermore, using an amino HPLC column, the aromatics were
pre-separated by ring number. The oil contained 8.2 % naphthalenes (all alkylated, starting
from the C2 species) as well as 1.8 % fluorenes, 3.1 % dibenzothiophenes, 2.2 %
anthracenes and phenanthrenes, 1 % fluoranthenes and pyrenes, 0.6 % chrysenes and 60
mg/kg benzopyrenes, all with at most 1 % non-alkylated parent compound (Grob et al.,
1991b). This should have underlined that batching oil was anything but a “white” oil. In fact, it

t
had been shown before to be a tumor promoter on mouse skin (Mehrotra et al., 1987).

ip
As the situation did not improve over the following years, another attempt was made to
increase pressure through scientific information: MOAH were analyzed in food. To this end,

cr
the probably most complex ever on-line HPLC-HPLC-GC system was built. Food extracts
were injected onto a large HPLC column, the fraction of the aromatics on-line reconcentrated

us
in a small cartridge filled with deactivated silica gel (solvent evaporator), then eluted onto a
smaller amino HPLC column, fractionated by ring number and transferred to GC-FID (Moret
an
et al., 1997). Now the Association of Chocolate, Biscuits and Confectionery Industries of
Europe (Caobisco), led by Nestlé, imposed a standard in agreement with the International
Jute Organisation (IJO): absence of mineral batching oil should be checked through the non-
M
saponifiable matter of an extract of the bag (a method feasible without sophisticated
instrumentation; IJO 1998). A limit of 1250 mg/kg jute was provided, taking into account the
non-saponifiable matter of vegetable oils used for batching.
ed

The contamination of foods arriving in Europe dropped, but did not disappear. In 2000, 30
samples of cocoa butter and cocoa masses sampled in Switzerland contained between <2
pt

and 73 mg/kg batching oil, with a mean of 20 mg/kg. In 2001, concentrations in 10 samples
of chocolate were all below 10 mg/kg, partly because of tight control, but batching oil
ce

remained detectable at a limit of 2 mg/kg in almost all samples. Some bag manufacturers
used both vegetable and mineral batching oils, depending on orders, resulting in cross
contamination in the machinery. With the specification of the non-saponifiable matter, some
Ac

contamination remained undetectable. Further, as for Asian rice, farmers seemed still to use
unsuitable bags to bring their produce to the trader.

Coatings and glazing agents

In 1990, it was detected that most rice on the Swiss market was sprayed with “white”
mineral oil to give it back the gloss lost by raffination. Concentrations were in the range of
800-2000 mg/kg rice, averaging about 1500 mg/kg (Grob et al., 1992). For an average
Swiss, yearly consuming approximately 4 kg rice, exposure to MOH from this source alone
was around 6 g/year. This practice was immediately stopped. It came as a surprise to find

7
rice on the Austrian market, supplied from Italy, that contained 1800 mg/kg MOH applied for
the same purpose still in 2008.

Heimbach et al. (2002) estimated dietary human exposure to waxes in the US to 0.020
mg/kg bw/d (around 0.45 g/person/year), with fruit and vegetable coating being the main
source. This was probably common practice for some fruits and vegetables, but there was
little data. Lawrence and Iyengar (1983) found MOH on squash, turnip, cucumber, sweet
potatoes, egg plants and most citrus fruits. Tennant (2004) considered confectionery glazing
agents as the main source of MOH for pre-school children.

t
This only partially applied to Switzerland. In the early nineteen nineties, no fruits or

ip
vegetables were found that had been coated with mineral waxes (those with a non-edible

cr
rind not being checked), nor were there indications that coatings were applied. MOH on dry
fruits, like raisins, were undetectable or at low levels, indicating contamination rather than

us
application to achieve an effect. Fiorini et al. (2010) reported that among 18 samples of dried
fruit (apricots, plums, raisins, coconut, dates, mangos and pineapples) two samples of
apricots contained 23 and 28 mg/kg MOH, respectively, forming humps of highly isomerized
an
material (i.e. indicating oils rather than waxes). It is questionable that this was applied
intentionally. Apparently coatings are still sometimes applied to spices: pepper with 300-400
mg/kg MOH was detected recently.
M

Release agents
Mineral oil products were widely used as release agents, mainly in the production of
ed

baking ware and sugar products. In industrial bakeries, they were used to spray all sort of
surfaces, e.g., of funnels through which dough had to glide, cutting devices to portion dough,
pt

to facilitate bread and biscuits leaving the mold, but also to cut loaves (toast bread) after
baking. In 1993, among 15 samples of toast bread and rusk, five contained less than 5 mg/kg
ce

MOH, two 65 and 190 mg/kg, respectively, and the remaining nine 600-2000 mg/kg MOH
(average 910 mg/kg). Among 12 samples of normal bread (uncut loafs), 5 from industrial
bakeries contained 35-180 mg/kg (average 85 mg/kg), suggesting that the highest levels
Ac

were from cutting the loaves (while still hot, to avoid microbial contamination). However, in
1995, among 20 samples of uncut bread, three contained 240, 470 and 1740 mg/kg MOH,
pointing out that other applications were also sometimes important.

With a mean yearly bread consumption of around 25 kg, as typical for many European
countries, and assuming regular consumption of toast bread containing 1000 mg/kg MOH,
exposure amounted to 25 g/y or almost 70 mg/d and was probably the major source of MOH
exposure. Interestingly, granulomas in human livers were most frequently found in North
America, Australia, New Zealand and Britain (Cruickshank, 1984), which might correlate with
the consumption of toast bread.

8
 In Switzerland, the probably last industrially used molds sprayed with a mineral oil release
agent (510 mg/kg in rusk) were replaced by a PTFE-coated mold in 1995. In 1999, all three
samples of rusk and the 16 samples of bread from Swiss industrial production analyzed
contained less than 5 mg/kg MOH. However, still four of ten samples of imported toast bread
contained 50, 70, 82 and 750 mg/kg MOH (centered on C20-C23, typical for release agents).
As an exceptional case, in 2003, an organic rye bread from Swiss production contained 1820
mg/kg MOH.

The use of specified white mineral oils as release agents or lubricant is still authorized by

t
the US-FDA up to 1500 mg/kg bakery product (FDA, 2017).

ip
Candies and similar confectionary frequently contained MOH. In 1990, of 24 bonbons

cr
eight contained 50-1300 mg/kg MOH, forming a chromatographic hump. For two types of
bonbons containing more than 1000 mg/kg MOH, it was shown that only about 5 mg/kg MOH

us
were extractable from the surface, indicating that the oil was distributed throughout the
bonbon and used as release agent before shaping (Grob et al., 1991c). In fact, throughout
the nineteen nineties, the reference oil used in the KLZ for measuring MOH was a release
an
agent from a candy producer. In 1993, the average concentration in 19 samples was 24
mg/kg MOH, with a maximum of 56 mg/kg, but there were also 6 samples with 250-1750
mg/kg MOH (mean of 850 mg/kg). These release agents were commonly centered at n-C21
M
to n-C23. In 1995, 5 samples of small Easter eggs made of colored sugar contained 110-
1300 mg/kg MOH, averaging 300 mg/kg.
ed

Some candies were packed in wax paper and contained 50-100 mg/kg MOH largely
consisting of n-alkanes, with an extreme product containing 2300 mg/kg wax. Chewing gums
usually contained 1-5 % wax-type MOH, but these are unlikely to be resorbed even if the
pt

chewing gum is finally swallowed.

ce

Lubricating oils
Lubricating oils are frequently named as expected food contaminants, but were rarely
encountered by analysis. They commonly consist of de-paraffinated mineral oil fractions with
Ac

a maximum abundance in the range of n-C23 to n-C28. In 1995, a stirring unit in machinery
for preparing butter lost lubricating oil, but the contamination remained undetectable (less
than the background contamination). A similar case occurred in potato chip manufacturing,
but the lost lubricating oil was again too diluted in a large amount of product to significantly
increase the MOH content.

In 1996, roughly 25 mg/kg of a mineral oil was detected in a cream. As rather usual, the
investigation at the quality control of the company was unsuccessful, but later it turned out
that the person at the dosing machinery had solved a problem without notifying his hierarchy:
the piston dosing the cream was lubricated, and to avoid that cream would flow backwards

9
and cause problems of hygiene, he increased the pressure of the lubricant to prompt oil
flowing forward into the cream rather than cream backward behind the piston. He
replenished the reservoir with Vaseline he bought in the pharmacy.

In 1997, 90 samples of canned oily foods, primarily fish in vegetable oil, were analyzed for
MOH. Concentrations in the vegetable oil were mostly around 100 mg/kg, with a maximum at
820 mg/kg (Grob et al., 1997). As sea foods are commonly contaminated by MOH (see
below), the fatty extract of the sea food and the added vegetable oil were analyzed
separately, from which it was concluded that the major amount of MOH must have originated

t
from the can, lubricating oil being suspected.

ip
Migration from food contact materials

cr
Plastics

us
EU legislation on plastics used as food contact materials (FCMs; Regulation (EU)
10/2011) authorizes the use of white mineral oils and waxes specified by an elevated
molecular mass (according to JECFA) without a specific migration limit (SML), i.e. up to the
an
overall migration limit of 10 mg/dm2 food contact surface.

Castle et al. (1991) found that migration of MOH from polystyrene cups and glasses into
M
beverages was below 0.5 mg/kg. In 1993, the same authors reported that the migration from
polystyrene containing up to 5.5 % MOH into cream and yoghurt was less than 1 mg/kg, but
migration into oil reached up to 140 mg/kg. Jickells et al. (1995) investigated a number of
ed

special applications and found that migration from cork lubricated with mineral hydrocarbons
into wine or from paper jam cover disks into preserves did not significantly exceed 1 mg/kg.
Migration from polystyrene and ABS containers containing 1-4 % MOH mostly amounted to
pt

<5-15 mg/kg, with exceptional higher values.

ce

There is little data on MOSH migration from polyolefins containing white mineral oil as
additive. Migration into powdered formula for baby bottles reached around 5 mg/kg
(Biedermann-Brem et al., 2012), which demonstrated that the migration of virtually non-
Ac

volatile substances into dry food may be significant when particles are small (Eicher et al.,
2015). There is no data on the migration into wetting, particularly fatty foods.

Paper and board

In 1993, three samples of powdered infant formula for baby bottles packed in paperboard
boxes and internal paper bags contained 8, 10 and 15 mg/kg MOH. Three years later, such
products were investigated more systematically and found to contain up to 58 mg/kg MOH,
depending on the duration of storage and whether the boxes were standing alone, exposed
to the atmosphere, or stacked on pallets or in the shelves of stores (Droz and Grob, 1997). In

10
1997, two samples reached 53 and 80 mg/kg MOH. The inks for printing the boxes were
assumed as a source, but recycled paperboard may have also contributed. Over the
following years, most producers of such infant products reacted by using an aluminum foil in
the internal bag that acts as a functional barrier.

It was soon recognized that recycled paperboard itself contained MOH (largely also from
printing inks, e.g. on newspaper). The pattern was virtually the same as from printing the box
and concentrations also typically ranged between 10 and 80 mg/kg dry foods kept for
extended periods at ambient temperature. In the early years 2000, the Swiss packers of rice

t
and a few other foods switched to fresh fiber board. This was, however, rather the exception:

ip
It was also widely argued that “the other cause” was responsible for such contamination,
namely that changing ink composition was little effective as long as recycled board was used

cr
and vice versa. For imported products, the common answer was that it is not possible to
make special packaging just for Switzerland. More plausible was the argument that cooking

us
of, e.g., rice or pasta in boiling water would eliminate the MOH – they are volatile, since they
were transferred to the food through the gas phase. However, an investigation concluded
an
that this was the case only for a minor part of the MOH, presumably because water acts a
barrier encapsulating the MOH in the pores of the particles (Biedermann-Brem and Grob,
2011).
M
In Switzerland, the use of mineral oil products as solvents for printing inks was banned
through the printing ink regulation from 2005 (FOPH, 2005). It resulted in the production of
ed

“low migration inks”, but imported products in particular often did not comply.

The situation changed when in December 2009 the German BfR stated that the migration
of mineral oil from boxes of recycled paperboard into dry foods should urgently be minimized
pt

(BfR, 2009). The German Federal Ministry of Food, Agriculture and Consumer Protection
(BMELV) immediately drafted a regulation on MOSH and MOAH migration from recycled
ce

paperboard. In parallel a printing ink regulation was developed that banned the mineral oil
from inks in food applications. In 2018, neither was yet in force. For the mineral oil regulation,
several problems remained: a conclusive toxicological evaluation is missing to support limit
Ac

values; the MOH are anyway just one of many non-evaluated substances migrating from
recycled paperboard (Biedermann and Grob, 2013), and MOH get into food also from many
other sources not covered by the regulation.

Folding boxes were in the main focus, but the largest amount of recycled fibers is used in
transport boxes. These may not only contaminate foods in paperboard folding boxes, but
also those in paper or plastic bags. It was shown for pasta that migration from transport
boxes through a plastic film and a box of fresh fiber board was substantial (Biedermann et
al., 2011). There were, however, also particular cases, like pralines or chocolate candies that
contained around 30 mg/kg MOH. After lengthy investigations of potential sources, it turned

11
out that the various types were prepared in larger quantities and stored on sheets of recycled
paperboard before being assorted in a collection.

Edible oils
It was soon recognized that edible oils were among the most seriously MOH-
contaminated foods. In Grob et al. (1994), results were reported for a pumpkin kernel oil (250
mg/kg MOH), a linseed oil (450 mg/kg), a safflower oil (500 mg/kg) and a soybean oil (1000
mg/kg). Numerous other oils exceeded 1000 mg/kg, particularly specialty oils like safflower

t
oils, nut oils, sesame oils or any cold pressed oils.

ip
In 1997, using the HPLC-HPLC-GC method with the on-line intermediate solvent
evaporation mentioned above, it was determined that in a safflower oil containing 2100

cr
mg/kg saturated MOH (later termed MOSH) hardly any aromatics (MOAH) were detectable
(Moret et al., 1997). This was confirmed for a number of other oils with high MOH contents

us
and suggested the use of while oils in the manufacture. This contamination soon
disappeared.
an
Seeds for specialty oils and copra were often delivered to the oil mill in jute bags,
contaminating the resulting oils at up to a few hundred mg/kg MOH.

In 2000, among 21 used frying oils, 9 contained less than 5 mg/kg MOH, 12 others a
M

mean of 55 mg/kg (maximum of 140 mg/kg) and as single one of 1660 mg/kg. The source
remained unknown, but it was assumed that part had been extracted from the fried products.
ed

In 2001, MOH concentrations in a broad range of commercial vegetable oils were reported
by Wagner et al. (2001). Now only few samples exceeded 100 mg/kg MOH. Moret et al.
(2009) and Tranchida et al. (2011) reported similar data. Peanut oils from Argentina were
pt

consistently contaminated at an average of 35 mg/kg MOH (maximum, 85 mg/kg; 52

samples). Concentrations in African peanut oils were substantially lower. In 2003, two wheat
ce

germ oils contained 840 and 1000 mg/kg MOH centered on n-C22.

The MOH contamination of edible oils may occur through the use of installations, tanks
Ac

and ships which had previously been used for mineral oil products. The contamination of
Ukrainian sunflower oil in 2008 is an example, with MOH concentrations reaching several
thousand mg/kg, of which deodoration removed part (though possibly into a fraction then
added to feed – see below).

In extra virgin olive oils, MOH concentrations were usually around 1 mg/kg, which is
thought to be an environmental contamination. In pressed or centrifuged olive oils of lower
quality (“virgin” or “lampant” oils), they were commonly in the range of a few mg/kg. However,
olive pomace oils (”sansa oils”; oils solvent-extracted or centrifuged from pomace) were (and
still are) typically contaminated at 100-400 mg/kg MOH. MOAH concentrations are usually

12
below 5 mg/kg. Several attempts were made to explain this (Wagner et al., 2001 and
unpublished data). The MOH seem to be largely located in the skin of the olives. As this skin
is poorly extracted in the production of extra virgin oil (well known for waxes and sterol
esters; e.g. Artho et al., 1993), also contaminants largely remain in the pomace. Lower
quality olive oils tend to contain more wax and sterol esters, presumably because of better
extractability of softer skins, which also applies to MOH. The bulk of the MOH, however,
remains in the residue and is extracted together with the relatively small amount of residual
oil. This is a plausible explanation for increased MOH contents in pomace oils, but it seems

t
difficult to explain the high levels encountered. In a thorough investigation, Moret et al. (2003)

ip
were unable to reach a firm conclusion, but had indications that the contamination increased
during storage of the pomace (often over several months under open atmosphere) and that

cr
mechanized handling, could explain the high levels. In 2003, 9 oils extracted from fresh
pomace in the laboratory contained 17-300 mg/kg MOH, with an average of 120 mg/kg,

us
which did not support this explanation. Gomez-Coca et al. (2016) suggested a contribution
by white mineral oils used in oil extraction machinery and talc typically containing around 25
an
mg/kg MOH (added to the pomace at about 2 %).

It was repeatedly hypothesized that the hydrocarbons claimed to be MOH were actually
formed in the plant material, e.g. by a fermentation process. It seems highly unlikely that
M
such complex mixtures of branched and cyclic hydrocarbons could be formed by enzyme-
supported processes and, in addition, Populin et al. (2004) demonstrated the presence of
hopanes (considered as markers for mineral oil) in relative concentrations similar to those in
ed

mineral oil products. Further, pomace was allowed to ferment in the laboratory over several
months and no increase in the hydrocarbons was found.
pt

Foods from animal sources

ce

Animals digest feeds contaminated from the environment, from contaminated fats and oils
added to the feed and from mineral oils intentionally added, e.g. as binders. Part of these
MOH end up in meat, milk or eggs.
Ac

Waste mineral oils in used edible oils added to feed

Early in 1999, it became known that in Belgium PCBs containing dioxins were dumped
into used edible oils added to animal feeds, resulting in severe poisoning of hens. This
prompted the control of the used edible oils (mostly frying oils) collected at public sites in
Switzerland, since these were also fed to animals. Tanks at these collection sites have two
entrances, on one side for edible oils and, on the other, for other liquid wastes, such as used
motor oils and unused pesticide preparations. It turned out that the liquids were not always
correctly discharged, but also that some restaurants collected their used frying oil in canisters

13
for mineral oil and/or together with used mineral oils. MOH concentrations varied widely.
Among 30 samples, two reached 7 and 25 %, eight others between 200 and 1000 mg/kg
(Grob et al., 2001). Forty samples from Austria gave similar results, with maxima of 1.4 and
10 % MOH.

In the feeds to which such oils had been added, concentrations often exceeded 100
mg/kg. In 58 samples of body fat of cows and pigs, the mean concentration was 25 mg/kg,
with a maximum of 150 mg/kg. Most of 20 samples of body fat of chickens contained less
than 5 mg/kg MOH, but there was also one with 150 mg/kg (Grob et al., 2001). The feeds

t
consumed by these animals was not available and, hence, no direct correlation was possible,

ip
but the situation improved when the admixture to feeds was stopped.

cr
In fact, in June 1999, it was decided to dispose of the edible oils from public collection
sites in the same way as are the mineral oils, mainly in the cement industry. Some attempts

us
were made to use them to feed the biological treatment of wastewater plants. A 30 mg
MOH/kg was applied. As this limit was exceeded too frequently, this was soon stopped.
an
Contaminated by-products of edible oil refining

However, lipids admixed to feeds remained an important source for feed contamination.
Among 18 fats and oils analyzed in 2002, 13 contained less than 20 mg/kg MOH, the others
M
20, 25, 35 and 1600 mg/kg MOH (the latter centered on n-C35, i.e. beyond the most
commonly used mineral oil products, such as lubricating oils). Hydrogenated fish oils ranged
from 80 to 600 mg/kg. Cottonseed oil for feed contained 300 mg/kg MOH (also cottonseed oil
ed

for human consumption; Biedermann et al., 2009). Four samples of fat for pigs from
Germany contained 145, 150, 200 and 375 mg/kg MOH.
pt

Physical refining of edible oils (today the usual procedure) strips the free fatty acids in the
deodoration process by water vapor at 220-260 °C. The recondensed acids were added to
ce

feeds, e.g. to milk for calves the fat of which had previously been removed to be sold as
butter or cream. Such deodoration also removes other volatile components, including MOH
up to roughly n-C25 (Wagner et al., 2001).
Ac

In 2001, among 28 samples of such refinery by-products (2 from Germany, the others
from Switzerland), none contained less than 120 mg/kg MOH; the average was 650 mg/kg,
the maximum 6800 mg/kg (Wagner et al., 2001). In the condensate from chemical refining
(extraction of the fatty acids by alkali previous to deodoration), as much as 10 g/kg MOH
were measured, since the MOH are reconcentrated into a far smaller condensate.

14
Binders for powders added to feeds

The analysis of feeds around the year 2000 revealed another source of MOH: binders
used for adding fines like vitamins and minerals to feed in order to prevent sedimentation in
sacks or silos. The concentrations in the feeds for hens varied between 300 and 1000 mg/kg.

The mean MOH concentration in the fat of eggs (40 samples analyzed in 2001) was 30
mg/kg, with a maximum of 80 mg/kg. Nearly all eggs were contaminated, including those
imported from many European countries. Eight eggs from Japan contained on average 22
mg/kg MOH, with a maximum of 35 mg/kg, suggesting that this was a widespread practice.

t
ip
In three experiments with feeds containing 400, 600 and 800 mg/kg MOH, it was
determined that the MOH concentrations in the fat of eggs corresponded to 1.5-3.0 % of that

cr
in the feed. As these were free range hens which ingested material other than these feeds,
these percentages were lower bound values. In 2008, 25 samples of Swiss and important

us
eggs were no longer contaminated in this way (less than 3 mg/kg MOH).

Milking balm
an
In 1995, the effect of using milking balm (purely MOH) was tested by a farmer intentionally
applying more milking balm than he would normally do for hand milking. Nonetheless, the
resulting milk contained less than 10 mg/kg MOH related to the fat and there was no
M

significant increase compared to the milk he previously obtained from the same cow.

MOH in foods of animal origin

ed

Among 10 samples of hard cheese analyzed in 1995, there was one containing 26 mg/kg
MOH related to the fresh weight, the others 10 mg/kg or less. Seven samples of butter
pt

contained between 12 and 37 mg/kg MOH, averaging 24 mg/kg. In 1999, 22 samples of

butter contained between 3 and 12 mg/kg MOH. In 2003, butter sold for cooking contained
ce

110 mg/kg hydrocarbons. It turned out that it had been made from butter having exceeded
the shelf life and melted out of the original packing. The hydrocarbons formed a hump of
unresolved components like MOH, but they might have consisted of oligomers of polyolefins.
Ac

In 2000, 10 samples of salami contained <5 to 20 mg/kg MOH (average of 13 mg/kg). Of

95 samples of meat (beef and pork) directly from the slaughterhouse, 43 samples contained
less than 5 mg/kg MOH, the others up to 55 mg/kg, with an average of 18 mg/kg.

Dust binders
Handling products like grain, soy beans or rice produces a dust hazard which may cause
explosions, which is why dust binders may be used. In the US, the use of specified white
mineral oil products is permitted as dust control agent for wheat, corn, soybean, barley, rice,

15
rye, oats, and sorghum at concentrations up to 200 mg/kg by weight of grain and at up to 800
mg/kg for rice (FDA, 2017). There is no such approval in Europe.

Numerous attempts were made to detect dust binders in products imported from North
America, but no MOH could be clearly attributed to them in rice or soy bean oil. In 1994,
among 50 samples of pasta, often made from American durum wheat, MOH concentrations
were below the detection limit of 2 mg/kg in 28 samples, whereas the average of the others
was 9 mg/kg, with a maximum of 19 mg/kg. More recently, 1.5-3 mg/kg MOSH (no MOAH)
were repeatedly detected in flour or pasta. The question remained open whether these were

t
residues of dust binders (the bulk possibly being removed with the husk and bran).

ip
Pesticide formulations

cr
Mineral oil is used to treat bark in winter against insects overwintering there. It appears
that pesticide formulations may also include mineral oils.

us
Pesticides were taken into consideration as source for the consistently high MOH
contamination of grapeseed oils (Fiorini et al., 2008). These oils are obtained from the
an
pomace of wine making and usually after fermentation to obtain distillates like grappa or
marc, i.e. from residues of residues. In 2000, 13 samples of grapeseed oils contained on
average 75 mg/kg MOH: 4 refined oils with 35-60 mg/kg; the oils with the highest
M
concentrations, 150 and 180 mg/kg MOH, were crude, the difference probably explained by
removal during deodorization. The main Swiss trader agreed to select products respecting a
ed

maximum of 50 mg/kg MOH, which must have required selection in its implementation. The
oil producers committed not to use mineral oil products.

In 2007, 11 commercial grape seed oils, Swiss and Italian, contained 43-247 mg/kg MOH
pt

(Fiorini et al., 2008). Deeper investigations revealed that the contamination was at least
predominantly from the residues of peels in the seeds. Oil obtained from the seeds of grapes
ce

grown by one of the authors without phytochemical treatment contained clearly less MOH (up
to 14 mg/kg) and the peels were less contaminated. This suggested an environmental
background contamination to which an additional contamination is added, but no firm
Ac

conclusions could be reached.

Fish and sea food

The fat extracts of 40 fishes, randomly collected from the Swiss market in 1996, contained
20-800 mg/kg MOH (mean of 350 mg/kg; 3-150 mg/kg referring to the whole fish). In the
widely scattering data, there was no obvious difference between salt and fresh water fish.
The MOH included aromatics, with 11 % naphthalenes but little 3-ring aromatics (Moret et al.,
1997).

16
 In 1999, of 6 locally caught pike, four contained less than 50 mg/kg MOH in the fat, the
others 310 and 360 mg/kg. Two trout contained 250 and 480 mg MOH/kg fat. Among 9
samples of sea fish, three contained less than 20 mg MOH/kg fat, the others on average 530
mg/kg. A fish containing 1060 mg MOH/kg fat was from The Netherlands, one with 1650
mg/kg from New Zealand. Calculated for the whole fish, these maxima corresponded to
about 10 and 30 mg/kg.

It was not possible to correlate MOH in fish with oil spills. A data set on background
contamination would be needed to check the impact of oil accidents. Commonly polycyclic

t
aromatic hydrocarbons (PAH) were analyzed (e.g. after the disaster in the Gulf of Mexico),

ip
but this is of questionable significance, since mineral oils contain almost exclusively alkylated
PAH (MOAH), and these do not belong to the strongly retained and accumulated

cr
hydrocarbons in animal or human tissues (Barp et al., 2014; Barp et al., 2017a).

us
Harvesting machinery
Background contamination of sunflower oils was investigated through seeds collected in
an
the area of Zurich (Fiselier and Grob, 2009; Grundboeck et al., 2010). MOSH concentrations
in the oils extracted with hexane from hand-picked sunflower seeds ranged from 0.1 mg/kg in
more rural area to 2.4 mg/kg in a field close to Zurich. In 7 samples of mechanically
M
harvested seeds, the concentrations were always higher, with an average of 5.6 mg/kg and a
maximum of 13.7 mg/kg. In some cases, diesel oil dominated, in others the lubricating oil,
perhaps depending on the temperature of the engine of the harvester. In two cases, seeds
ed

dried at the collection site contained 30 and 41 mg MOSH/kg oil, dominantly diesel oil,
probably because the air sucked through the seeds was contaminated by delivering and
pt

unloading tractors.

In 32 commercial refined sunflower oils from the Italian and Swiss market, the MOSH
ce

concentrations varied between 2.7 and 32 mg/kg, with an average of 11.2 mg/kg. Since
deodorization must have removed most of the MOH up to around n-C25, either the
concentrations in the crude oils was even higher or there was an additional contamination in
Ac

the oil mill.

Wheat germs analyzed in 2001 contained 250 mg MOH/kg extracted oil. Fifteen
commercial oils contained on average 65 mg/kg MOH, none less than 20 mg/kg. In 24
samples of wheat, 0.4-5 mg/kg MOH were measured (average of 2.8 mg/kg), which
corresponded to 30-340 mg/kg in the oil (fat content, 1-2 %). In 2003, oil extracted in the
laboratory from 10 samples of grain contained 48-400 mg/kg MOH (average of 135 mg/kg). It
was assumed that this was again a combination of environmental contamination and an
addition, e.g., by harvesters. Concentrations were high because of strong enrichment into a
low content of lipids.

17
 Environmental contamination

MOH in the environment

As mentioned above, part of the food contamination with MOH is from the environment. In
2002, air was analyzed for the mass range typically found in food, i.e. C20-50. As these
MOH are largely in the particulate matter, they were determined in the dust sampled by filters
(PM10). In a road tunnel, around 5 µg MOH/m3 air were measured, corresponding to about 3
% of the PM10 (Neukom et al., 2002). The molecular mass distribution corresponded to that

t
found in the particulate matter from the exhaust of hot diesel engines (Brandenberger et al.,

ip
2005), which in turn largely corresponded to that of engine (lubricating) oil. In Swiss cities,
the MOH amounted to 0.1-1.5 µg/m3 (about 1 % of the dust), in rural areas to around 0.03

cr
µg/m3 (0.3% of the dust). They were mainly in the range of C25-35, the low mases probably
from incompletely burned heating or diesel oil (soot), the center at C27-29 from lubricating oil

us
and the high masses (sometimes up to C50) maybe from debris of road tar.

Humus in a garden in the town of Zurich contained slightly less than 1 mg/kg MOH, in a
an
more rural area around 0.5 mg/kg. Below trees, the MOH concentrations were several times
higher; below a fir tree in the forest it was 7.5 mg/kg. In compost it reached 38 mg/kg
(Neukom et al., 2002), indicating accumulation in rotting plant material.
M
Soil from the natural floor of a cellar from the early 19th century contained 0.1 mg/kg MOH
at the surface and less than 0.02 mg/kg 5 cm below it, confirming that the hydrocarbons are
ed

from the modern environmental contamination.

Uptake into plant material

pt

Plants grow in a MOH-contaminated environment, with deposition from the air and uptake
from the soil. Animals fed contaminated plant material partly digest the MOH and transfer the
ce

others to our food. No literature on the uptake of MOSH or MOAH into plants was found, but
much work on the uptake of polycyclic aromatic hydrocarbons (PAH) directly from the air and
via the soil (e.g. Kipopoulou et al., 1999; Parrish et al., 2006; Lin et al., 2007; Zezulka et al.,
Ac

2014; Doucette et al., 2018; Zhan et al., 2018).

Some experiments were performed on the uptake of MOH. Such analyses were
demanding because of low concentrations, high risks of contamination during sample
preparation and the often required prior removal of the natural n-alkanes (by activated
aluminum oxide; Fiselier et al., 2009).

MOH concentrations were determined in leaves of a beech from their appearance (May 8,
2001) to freshly dropped ones (November 12). Concentrations referred to the dry weight
were almost constant (Neukom et al., 2002), which would not be expected from deposition of
dust onto the leaves.

18
 In 2003, 6 samples of lettuce contained 0.06-0.35 mg/kg MOH per dry mass, other 5
samples 0.35, 0.7, 1.2, 1.5 and 7 mg/kg. For the two with the highest contamination, no
significant difference was measured for internal (completely protected) and external leaves.
For lettuce analyzed later, the internal leaves contained 0.6 mg/kg MOH, the external ones
0.9 mg/kg dry matter. Also radish grown below the soil surface contained MOH: 1.2 mg/kg
dry mass in the center, 7 mg/kg in the skin. The green part above ground contained 1.4 mg
MOH/kg dry mass. This suggests that most MOH (presumably largely MOSH) are picked up
via contaminated soil. The finding that most MOH in olives are located in the skin would have

t
to be explained by the flow of water evaporating from the surface. No doubt, major questions

ip
still await answers.

cr
Modified MOH composition

The composition of the MOH in the plants, e.g. the vegetables and fruits humans

us
consume, is not the same as that of the mineral oil products released into the atmosphere. In
the plant material, the MOAH are strongly reduced (often almost absent), probably by
oxidation in the air (e.g. Nikolaou et al., 1984; Molteni et al., 2018). Also some MOSH might
an
be oxidized. The easily digestible MOSH are then converted by microorganisms in the soil
and perhaps also in plants, reconcentrating the species which might also be accumulated by
M
humans.

Animal feeds are contaminated by the same environmental MOH as the vegetables
consumed by humans. Assuming a MOSH concentration in grass similar to that in lettuce or
ed

leaves of trees (see above), a cow daily feeding around 20 kg hay (or dry mass of grass)
might ingest 50-100 mg MOSH/d. Part of these MOSH was absorbed and metabolized,
pt

another part accumulated in what becomes meat (for rats shown by Barp et al., 2017a) or
excreted into milk. Milk seems typically to contain around 10 mg MOSH/kg fat. If a cow daily
produces 20 L milk with 3.5 % fat, it excretes 7 mg MOH/d, which is roughly 10 % of the
ce

ingested MOSH. Presumably the MOH in the milk are further modified in the direction of well
absorbed and poorly degradable constituents – towards the hydrocarbons that are strongly
Ac

accumulated by humans. The toxicological assessment may need to take this into account.

Outlook
This review covers the probably less known work on MOH in foods up to about 2010,
when MOH became a subject of broad interest. It was the period when high concentrations
were detected, but also the main sources of contamination eliminated, reduced or at least
recognized. Present exposure of consumers is substantially lower than it was in the nineteen
nineties.

19
 Most investigations reported above are incomplete. For each type of food, it remains to
determine the level of MOH from environmental sources, as this background contamination
will be difficult to avoid. Additional MOH contamination is, once known, more easily avoidable
in most cases. Some data suggests that for numerous foods the contamination is now no
longer far above the background.

In Europe, a broad monitoring program of the MOH concentrations in food is on-going

(European Commission, 2017) to enable an estimate of the present human exposure. This
estimate should then be evaluated against toxicological data.

t
The toxicological evaluation of the past needs revision (Grob, 2018). Accumulation in

ip
human tissues has been grossly underestimated. The high MOSH concentrations

cr
determined in human liver, spleen and lymph nodes reached levels similar to (in extreme
cases even exceeding) those in rats exposed to the highest experimental dose (Barp et al.,

us
2017a,b). There is insufficient information about relevant toxicological end points and
potential adverse effects at these levels. For mineral oils directly entering food, experimental
tests can be conducted with the given types of oil, but for MOH modified by light,
an
microorganisms, plants and/or animals, today possibly the dominant source, it might be
necessary to account for an enrichment of the accumulating MOSH. For the MOAH, the
composition is more relevant than the total concentration: whereas in some cases the MOAH
M
almost exclusively consist of mono- and diaromatics, in others they include (alkylated)
polyaromatics.
ed

Despite these important gaps, a revision of the TDIs falls due, since numerous
applications resulting in high concentrations in food are legally approved. MOSH
concentrations in human tissues are already high (Barp et al., 2014; EFSA, 2012), but
pt

present TDIs for MOSH of elevated molecular mass are above the present exposure by 2-3
orders of magnitude (JECFA, 2002) and authorize a corresponding increase of the exposure.
ce

This needs to be corrected.

References
Ac

Artho A, Grob K, Mariani C. 1993. On-Line LC-GC for the analysis of the minor components
in edible oils and fats - the direct method involving silylation. Fat Sci Techn. 95:176-
180.
Barp L, Kornauth C, Würger T, Rudas M, Biedermann M, Reiner A, Concin N, Grob K. 2014.
Mineral oil in human tissues, part I: concentrations and molecular mass distributions.
Food Chem Toxicol. 72:312-321.
Barp L, Biedermann M, Grob K, Blas-Y-Estrada F, Nygaard UC, Alexander J, Cravedi JP.
2017a. Accumulation of mineral oil saturated hydrocarbons (MOSH) in female Fischer

20
 344 rats: comparison with human data and consequences for risk assessment. Sci
Total Environ. 575:1263-1278.
Barp LM, Biedermann M, Grob K, Blas-Y-Estrada F, Nygaard UC, Alexander J, Cravedi JP.
2017b. Mineral oil saturated hydrocarbons (MOSH) in female Fischer 344 rats;
accumulation of wax components; implications for risk assessment. Sci Total Environ.
583:319–333.
BfR (2009) German Federal Institute for Risk Assessment (BfR). Übergänge von Mineralöl
aus Verpackungsmaterialien auf Lebensmittel. Stellungnahme Nr. 008/2010 des BfR

t
vom 09. Dezember 2009,

ip
http://www.bfr.bund.de/cm/216/uebergaenge_von_mineraloel_aus_verpackungsmate
rialien_auf_lebensmittel.pdf

cr
BfR (2015) Recommendation XXXVI. Papiere, Kartons und Pappen für den
Lebensmittelkontakt. http://bfr.zadi.de/kse/faces/resources/pdf/360.pdf

us
BfR (German Federal Institute for Risk Assessment) (2012) 10. Sitzung der BfR Kommission
für Bedarfsgegenstände, Protokoll vom 29. November 2012. p.
an
6http://www.bfr.bund.de/cm/343/10-sitzung-der-bfr-kommission-
fuerbedarfsgegenstaende.pdf
Biedermann M, Grob K, Meier W. 1989. Partially concurrent eluent evaporation with an early
M
vapor exit; detection of food irradiation through coupled LC-GC analysis of the fat. J
High Resol Chromatogr. 12:591-598.
Biedermann M, Grob K. 2009. How "white" was the mineral oil in the contaminated Ukrainian
ed

sunflower oils? Eur J Lipid Sci Technol. 111:313–319.

Biedermann M, Fiselier K, Grob K. 2009. Aromatic hydrocarbons of mineral oil origin in
foods: method for determining the total concentration and first results. J Agric Food
pt

Chem. 57:8711-8721.
Biedermann M, Ingenhoff JE, Barbanera M, Garbini D, Grob K. 2011. Migration of mineral oil
ce

into noodles from recycled fibers in the paperboard box and the corrugated board
transport box as well as from printing inks: a case study. Packag Techn Sci. 24:281-
Ac

290.
Biedermann M, Grob K. 2013. Assurance of safety of recycled paperboard for food
packaging through comprehensive analysis of potential migrants is unrealistic. J
Chromatogr A. 1293:107-119.
Biedermann M, Grob K. 2015. Comprehensive two-dimensional gas chromatography for
characterizing mineral oils in foods and distinguishing them from synthetic
hydrocarbons. J Chromatogr A. 1375:146–153.

21
Biedermann-Brem S, Grob K. 2011. Removal of mineral oil migrated from paperboard
packing during cooking of foods in boiling water. Eur Food Res Technol. 232:1035-
1041.
Biedermann-Brem S, Kasprick N, Simat T, Grob K. 2012. Migration of polyolefin oligomeric
saturated hydrocarbons (POSH) into food. Food Addit Contam A. 29:449-460.
Brandenberger S, Mohr M, Grob K, Neukom HP. 2005. Contribution of unburned lubricating
oil and diesel fuel to particulate emission from passenger cars. Atmos Environ. 39:
6985-6994.

t
Castle L, Kelly M, Gilbert J. 1991. Migration of mineral hydrocarbons into food. 1.

ip
Polystyrene container for hot and cold beverages. Food Addit Contam. 8:693-700.
Castle L, Kelly M, Gilbert J. 1993. Migration of mineral hydrocarbons into food. 2.

cr
Polystyrene. ABS, and waxed paperboard containers for dairy products. Food Addit
Contam. 10:167-174.

us
Lawrence JF, Iyengar JR. 1983. Determination of paraffin wax and mineral oil on fresh fruits
and vegetables by high temperature GC. J Food Safety. 5:119-129.
an
Cruickshank B. 1984. Follicular (mineral oil) lipidosis: I. Epidemiologic studies of involvement
of the spleen. Hum Pathol. 15:724-730.
Dincsoy HP, Weesner RE, MacGee J. 1982. Lipogranulomas in non-fatty human livers. A
M
mineral oil induced environmental disease. Am J Clin Pathol. 78, 35-41.
Doucette WJ, Shunthirasingham C, Dettenmaier EM, Zaleski RT, Fantke P, Arnot JA. 2018.
Review of measured bioaccumulation data on terrestrial plants for organic chemicals:
ed

metrics, variability, and the need for standardized measurement protocols. Environ
Toxicol Chem. 37:21-33.
Droz C, Grob K. 1997. Determination of food contamination by mineral oil material from
pt

printed cardboard using on-line coupled LC-GC-FID. Z Lebensm Unters Forsch.

205:239-241.
ce

EFSA. 2012. European Food Safety Authority, Scientific opinion on mineral oil hydrocarbons
in food. EFSA J. 10(6):2704, 1–185.
Ac

http://www.efsa.europa.eu/en/efsajournal/pub/2704.htm
EFSA. 2013. European Food Safety Authority, Scientific opinion on the safety assessment of
medium viscosity white mineral oils with a kinematic viscosity between 8.5–11 mm2/s
at 100 °C for the proposed uses as a food additive. EFSA J., 11, 3073.
http://www.efsa.europa.eu/de/efsajournal/pub/3073.htm.
Eicher A, Biedermann M, Zurfluh M, Grob K. 2015. Migration by “direct” or “indirect” food
contact? Dry or wetting foods? Some experimental data for paper and board. Food
Addit Contam A. 32:110-119.

22
European Commission (2017) Commission Recommendation (EU) 201/84 on the monitoring
of MOH in food and in materials and articles intended to come into contact with food.
http://eur-lex.europa.eu/legal-
content/EN/TXT/PDF/?uri=CELEX:32017H0084&from=EN
FDA (2017) CFR - Code of Federal Regulations Title 21. Part 172 – food additives permitted
for direct addition to food for human consumption.
https://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=172.878
Fiorini D, Fiselier K, Biedermann M, Ballini R, Coni E, Grob K. 2008. Contamination of grape

t
seed oil with mineral oil paraffins. J Agric Food Chem. 58:11245-11250.

ip
Fiorini D, Paciaroni A, Gigli F, Ballini R. 2010. A versatile splitless injection GC-FID method
for the determination of mineral oil paraffins in vegetable oils and dried fruit. Food

cr
Control. 21:1155–1160.
Fiselier K, Grob K. 2009. Determination of mineral oil paraffins in foods by on-line HPLC–

us
GC–FID: lowered detection limit; contamination of sunflower seeds and oils. Eur Food
Res Technol. 229:79–688.
an
Fiselier K, Fiorini D, Grob K. 2009. Activated aluminum oxide selectively retaining long chain
n-alkanes. Part II, integration into an on-line HPLC-LC-GC-FID method to remove
plant paraffins for the determination of mineral paraffins in foods and environmental
M
samples. Anal Chim Acta. 634:102-109.
FOPH. 2005. Federal Office of Public Health. Swiss regulation SR 817.023.21 Verordnung
des EDI über Bedarfsgegenstände vom 23. November 2005.
ed

http://www.admin.ch/ch/d/sr/817_023_21/index.html
Gomez-Coca R, Perez-Camino MC, Moreda W. 2016. Saturated hydrocarbon content in
olive fruits and crude olive pomace oils. Food Addit Contam A. 33:391-402.
pt

Grob, K. 2018. Toxicological assessment of mineral hydrocarbons in foods. J Agric Food

Chem. (submitted)
ce

Grob K, Fröhlich D, Schilling B, Neukom HP, Nägeli P. 1984. Coupling of HPLC with
Capillary GC. J Chromatogr. 295:55-61.
Ac

Grob K, Lanfranchi M, Egli J, Artho A. 1991a. Determination of food contamination by

mineral oil from jute bags using coupled LC-GC. J Assoc Off Anal Chem. 74:506-512.
Grob K, Biedermann M, Caramaschi A, Pacciarelli B. 1991b. LC-GC analysis of the
aromatics in a mineral oil fraction: batching oil for jute bags. J High Resol
Chromatogr. 14:3-39.
Grob K, Artho A, Biedermann M, Egli J. 1991. Food contamination by hydrocarbons from
lubricating oils and release agents; determination by coupled LC-GC. Food Addit
Contam. 8:437-446.

23
Grob K, Artho A, Biedermann M, Egli J, Lanfranchi M, Caramaschi A, Etter R, Romann E.
1992. Wieviel Mineralöl ist in den Lebensmitteln tolerierbar? Mitt Gebiete Lebensm
Hyg. 83:40-54.
Grob K, Biedermann M, Bronz M. 1994. Resultate einer Kontrolle von Speiseölen:
Fälschungen durch Zumischungen, Verunreinigungen. Mitt Gebiete Lebensm Hyg.
85:351-365.
Grob K, Huber M, Boderius U, Bronz, M. 1997. Mineral oil material in canned foods. Food
Addit Contam. 14:83-88.

t
Grob K, Vass M, Biedermann M, Neukom HP. 2001. Contamination of animal feed and food

ip
from animal origin with mineral oil hydrocarbons. Food Addit Contam. 18:1-10.
Grundböck F, Fiselier K, Schmid F, Grob K. 2010. Mineral oil in sunflower seeds: the

cr
sources. Eur Food Res Technol. 231:209–213.
Heimbach JT, Bodor AR, Douglass JS, Barraj LM, Cohen SC, Biles RW, Faust HR. 2002.

us
Dietary exposures to mineral hydrocarbons from food-use. Food Chem Toxicol. 40,
555–571.
an
IJO. 1998. International Jute Study Group, Food grade jute bag, IJO standard 98/1.
http://jute.org/jute_prod_fgjb.html
JECFA (Joint FAO/WHO Expert Committee on Food Additives) (2002) 59th report, 11–20;
M
WHO Technical Report Series 913. http://whqlibdoc.who.int/trs/WHO_TRS_913.pdf
Jickells SM, Nichol J, Castle L (1995) Migration of mineral hydrocarbons into food. 5.
Miscellaneous applications of mineral hydrocarbons in food contact materials. Food
ed

Addit Contam 11:333-341

Kipopoulou AM, Manoli E, Samar C. 1999. Bioconcentration of polycyclic aromatic
hydrocarbons in vegetables grown in an industrial area. Environ Pollut 106:369-380
pt

Lin H, Tao S. Zuo Q, Coveney RM (2007) Uptake of polycyclic aromatic hydrocarbons by

maize plants. Environ Pollution. 148: 614-619.
ce

Lorenzini R, Fiselier K, Biedermann M, Barbanera M, Braschi I, Grob K. 2010. Saturated and

aromatic mineral oil hydrocarbons from paperboard food packaging: estimation of
Ac

long-term migration from contents in the paperboard and data on boxes from the
market. Food Addit Contam A. 27:1765–1774.
Mehrotra NK, Kumar S, Agarwal R, Antony M. 1987. Jute batching oil: a tumor promoter on
mouse skin. Environ Res. 42:12–23.
Meudec A, Dussauze J, Deslandes E, Poupart N. 2006. Evidence for bioaccumulation of
PAHs within internal shoot tissues by a halophytic plant artificially exposed to
petroleum-polluted sediments. Chemosphere. 65:474-481.

24
Molten U, Bianchi F, Klein F, El Haddad I, Frege C, Rossi MJ, Dommen J, Baltensperger U.
2018. Formation of highly oxygenated organic molecules from aromatic compounds.
Atmos Chem Phys. 18:1909-1921.
Moret S, Grob K, Conte LS. 1997. Mineral oil polyaromatic hydrocarbons in foods, e.g. from
jute bags, by on-line LC-solvent evaporation (SE)-LC-GC-FID. Z Lebensm Unters
Forsch. 204:241-246.
Moret S, Populin T, Conte LS. 2009. Contamination of vegetable oils with mineral oils. Riv
Ital Sostanze Grasse. 86:3-14.

t
Neukom HP, Grob K, Biedermann M, Noti A. 2002. Food contamination by C20-C50 mineral

ip
paraffins from the atmosphere. Atmos Environ. 36:4839-4847.
Nikolaou K, Masclet P, Mouvier G. 1984. Sources and chemical reactivity of polynuclear

cr
aromatic hydrocarbons in the atmosphere – a critical review. Sci Tot Environ. 32:102-
132.

us
Noti A, Grob K, Biedermann M, Deiss U, Brüschweiler BJ. 2003. Exposure of babies to C15-
C45 mineral paraffins from human milk and breast salves. Regul Toxicol Pharm.
38:317-325.
an
Parrish ZD, White JC, Isleyen M, Gent MPN, Iannucci-Berger W, Eitzer BD, Kelsey JW,
Mattina MI. 2006. Accumulation of weathered polycyclic aromatic hydrocarbons
M
(PAHs) by plant and earthworm species. Chemosphere. 64:609-618.
Populin T, Biedermann M, Grob K, Moret S, Conte L. 2004. Relative hopane content
confirming the mineral origin of hydrocarbons contaminating foods and human milk.
ed

Food Addit Contam. 21:893-904.

SCF (1989) Scientific Committee on Food. Minutes of the 67th Meeting. III/3681/89-EN
SCF (1995) Scientific Committee on Food. Opinion expressed on 22 September, 1995
pt

(Annex 5 to Document III/5611/95)

Scotter MJ, Castle L, Massey RC, Brantom PG, Cunninghame ME. 2003. A study of the
ce

toxicity of five mineral hydrocarbon waxes and oils in the F344 rat, with histological
examination and mineral oil hydrocarbons in food tissue-specific chemical
Ac

characterisation of accumulated hydrocarbon material. Food Chem. Toxicol. 41:489-

521.
Tennant DR. 2004. The usage, occurrence and dietary intakes of white mineral oils and
waxes in Europe. Food Chem Toxicol. 42:481-492.
Tranchida PQ, Zoccali M, Purcaro G, Moret S, Conte L, Beccaria M, Dugo P, Mondello L.
2011. A rapid multidimensional liquid–gas chromatography method for the analysis of
mineral oil saturated hydrocarbons in vegetable oils. J Chromatogr A. 1218:7476-
7480.

25
Wagner C, Neukom HP, Grob K, Moret S, Populin T, Conte LS. 2001. Mineral paraffins in
vegetable oils and refinery by-products for animal feeds. Mitt Lebensm Hyg 92:499-
514.
Zezulka S, Klemš M, Kummerová M. 2014. Root and foliar uptake, translocation, and
distribution of [14C] fluoranthene in pea plants (Pisum sativum). Environ Toxicol
Chem. 33:2308-2312.
Zhan X, Zhu M, Shen Y, AYue L, Li J, Gardea-Torresdey JL, Xu G. 2018. Apoplastic and
symplastic uptake of phenanthrene in wheat roots. Environ Pollution. 233:331-339.

t
Zurfluh M, Biedermann M, Grob K. 2014. Enrichment for reducing the detection limits for the

ip
analysis of mineral oil in fatty foods. J Consum Prot Food Saf. 9:61-69.

cr
us
an
M
ed
pt
ce
Ac

26
 Figure 1
Ac
ce
pt
ed

27
M
an
us
cr
ip
t