BEU, Alm.del - 2021-22 - Bilag 356: Orientering om potentielt kræftfremkaldende stoffer i spraybaserede rense og smøremidler, fra beskæftigelsesministeren

Beskæftigelsesudvalget 2021-22
BEU Alm.del Bilag 356
Offentligt

Toxicology 477 (2022) 153261

Contents lists available at

ScienceDirect

Toxicology

journal homepage:

www.elsevier.com/locate/toxicol

Identification of substances with a carcinogenic potential in

spray-formulated engine/brake cleaners and lubricating products, available

in the European Union (EU)

–

based on IARC and EU-harmonised

classifications and QSAR predictions

Jorid B. Sørli

a, *

, Marie Frederiksen

, Nikolai G. Nikolov

, Eva B. Wedebye

, Niels Hadrup

a, c, **

National Research Centre for the Working Environment, DK-2100 Copenhagen, Denmark

DTU quantitative structure-activity relationships (QSAR) team, Research Group for Chemical Risk Assessment and GMO, National Food Institute, Technical University

of Denmark, Denmark

Division of Diet, Disease Prevention and Toxicology, National Food Institute, Technical University of Denmark, Denmark

A R T I C L E I N F O

Keywords:

Cancer

Greaser

Degreaser

Occupational exposure

Aerosol exposure

Hydrocarbon

A B S T R A C T

Spray-formulated engine/brake cleaners and lubricating agents are widely used to maintain machines. The

occupational exposure to their aerosols is evident. To assess the carcinogenic potential of these products, we

identified such products available in the European Union (EU). We built a database with CAS numbers of 1)

mono-constituent substances, and 2) multi-constituent-substances, and unknown-or-variable-composition,-

complex-reaction-products-and-biological-materials (multi-constituent/UVCBs). The compositions of multi-

constituent/UVCBs were unravelled with European Chemicals Agency (ECHA) registration dossiers. To iden-

tify carcinogenic potentials, we searched for 1) International Agency for Research on Cancer (IARC) classifica-

tion; 2) Harmonised classifications in Annex VI to the EU classification, labelling and packaging (CLP)

Regulation; and 3) whether they had a Danish Environmental Protection Agency advisory CLP self-classification

based on quantitative structure-activity relationships (QSARs) for genotoxicity and carcinogenicity in the Danish

(Q)SAR Database. In 82 products, we identified 332 mono-constituent substances and 44 multi-constituent/

UVCBs. Six substances were either IARC 1 or 2B classified. Twelve mono-constituent substances and 22 multi-

constituent/UVCBs had harmonised classifications as Carcinogenic Category 1A, 1B or 2, while nine sub-

stances fulfilled the QSAR-based advisory self-classification algorithms for mutagenicity or carcinogenicity. At

the product level, 39 products contained substances of carcinogenic concern by either IARC, harmonised clas-

sification or QSAR. We conclude that in the investigated EU marketed spray-formulated engine/brake cleaners

and lubricants, 24 of 332 mono-constituent substances and 28 of 44 multi-constituent/UVCBs had a carcinogenic

potential. At the product level, 39 of 82 contained substances with an identified carcinogenic potential. Regu-

lators and manufacturers can use this determination of carcinogenic potential to decrease occupational risk.

1. Introduction

Cancer is a major cause of death, and limiting carcinogenic risk is of

high concern. Workers employed in maintaining machines are exposed

to substances that may have carcinogenic potential, e.g. hydrocarbons

(IARC,

2012).

Individuals working with engines have a four times higher

risk for developing lung cancer compared to the general population,

according to a study of cancer in 15 million workers in the Nordic

countries (Pukkala

et al., 2009).

A study looking at the cancer risk for

Danish seafarers evaluated the risk for 33.000 people based on their

work title; seafarers compared to the general population had an

increased cancer risk of 1.3 (i.e. 30% increase) for men and 1.1 for

women. When the group was divided by work title, the engine-room

crew of ships had an increased risk of 2.3, and the maintenance crew

* Corresponding author.

** Corresponding author at: National Research Centre for the Working Environment, DK-2100 Copenhagen, Denmark.

E-mail addresses:

(J.B. Sørli),

(M. Frederiksen),

(N.G. Nikolov),

(E.B. Wedebye),

(N. Hadrup).

https://doi.org/10.1016/j.tox.2022.153261

Received 7 June 2022; Received in revised form 11 July 2022; Accepted 15 July 2022

Available online 18 July 2022

nc-nd/4.0/).

BEU, Alm.del - 2021-22 - Bilag 356: Orientering om potentielt kræftfremkaldende stoffer i spraybaserede rense og smøremidler, fra beskæftigelsesministeren

J.B. Sørli et al.

Toxicology 477 (2022) 153261

had an increased cancer risk of 4.2 (Kaerlev

et al., 2005; Ugelvig

Petersen et al., 2018).

Substances in machine maintenance shown to hold a carcinogenic

potential include trichloroethylene, a degreaser for metal, identified as

“carcinogenic

to humans” by IARC (IARC,

2014).

Moreover, oil products

as well as polycyclic aromatic hydrocarbons (PAHs), and exhaust gases

from diesel and gasoline were suggested as contributors to the increased

cancer risk of Danish seamen (Kaerlev

et al., 2005).

Mineral oils are

frequently used in engine cleaning and lubricant agents. ‘Mineral oils,

untreated or mildly treated’ were classified as a Group 1 carcinogen by

IARC (IARC,

2012),

i.e. there is enough evidence to conclude that it can

cause cancer in humans, while IARC back in 1984 wrote that ‘more

refined mineral oils’ is

“not

classifiable as to its carcinogenicity to

humans“ (i.e. Group 3) (IARC,

1984).

One thing is that these products may hold a carcinogenic potential.

Another is how high the exposure to the products is, as risk to the user is

the product of hazard (carcinogenic potency) and exposure. Besides

exposure via the skin, which may be substantial, inhalation is another

important exposure pathway with potential transfer over the large sur-

face area of the lungs. Thus, if the chemicals are formulated in spray

form, the inhalation of aerosols will likely result in significant exposure

for the workers. Such spray-formulated products for engine maintenance

include cleaning agents (degreasers) and lubricating agents.

To limit exposure to potentially carcinogenic substances in sprays,

we need to know their identity and the extent to which they are used. To

address this issue, we retrieved the CAS numbers of each individual

chemical substance by both covering mono-constituents and multi-

constituent/UVCBs. We first evaluated the list for IARC and EU-

harmonised classifications (Fig.

1);

we then supplemented with the

Danish Environmental Protection Agency (EPA) so-called advisory self-

classifications (The advisory list for self-classification of hazardous

substances) (EPA_Denmark,

2022; Wedebye et al., 2017),

genotoxicity

and carcinogenicity. The Danish EPA self-classifications are based on

predictions from the Danish (Q)SAR Database from several QSAR

models and following a weight-of-evidence (WoE) algorithm.

The justification for our application of these three approaches: 1)

IARC is an important authority of the World Health Organisation

(WHO); 2) Harmonised classifications is a system established in the EU

to ensure adequate risk management of substances with known muta-

genicity and/or carcinogenicity hazards, harmonised through classifi-

cation and labelling based on assessments and regulatory adoption

resulting in listing in Annex VI to the EU CLP Regulation. 3) Finally, we

applied QSAR to identify possible additional substances with predicted

carcinogenic potential based on the chemical structure. QSAR was used

as a supplement for substances that had not been identified in the former

two systems, e.g. if extensive animal tests or epidemiological studies had

not investigated their carcinogenic potential in humans.

2. Methods

2.1. Database of spray-formulated engine/brake cleaners and lubricating

substances

An overview of the work process is provided in

Fig. 1.

We searched

the internet for spray-formulated engine/brake cleaners and lubricants

that could be purchased in the EU (or European Economic Area: EAA)

using the Google search engine with combinations of the following

words in English and Danish (Danish words not provided here):

“greasing”, “greaser”,

degreasing”

“degreaser”, “engine

cleaner”,

“brake

cleaner”,

“wheel/rim

cleaner”, in combination with

“spray”.

The

searches were mainly performed during the time period 1 March to 31

May 2020. Using information in Safety Data Sheets (SDSs) available on

the internet, a list of CAS numbers was established in an Excel database

(Supplemental File 1). The amount of each mono-constituent or multi-

constituents/UVCBs in the products was noted only for substances

with a carcinogenic potential determined by the approaches outlined in

the next sections.

For multi-constituent substances/UVCBs, the CAS number of each

known individual constituent was identified by inspection in the section

“Compositions”

of the registration dossier found on the ECHA portal. For

example, the identity of the constituents in the multi-constituent/UVCB

with CAS number 64742–47–8 was found using the

“Compositions”

the Registration Dossier (https://echa.europa.eu/nl/registration-

dossier/-/registered-dossier/15375/1). The constituents that were part

of multi-constituent/UVCBs are also included in the Excel database

(Supplemental File 1). Notably, for some multi-constituent/UVCBs,

composition information was not available in ECHA registration dos-

siers, and thus their constituents could not be identified. For some

substances with multiple constituents, we could not retrieve information

in ECHA registration dossiers on whether they were defined as multi-

constituents or UVCBs. Yet, we still use the designation multi-

constituent/UVCB. The purpose of the project was to characterise the

overall extent of substances with a carcinogenic potential and not to

point to specific products. Therefore the product names are omitted in

this database. Instead, each product is given a number.

2.2. IARC classification and search for EU-harmonised classification as

provided in Annex VI to CLP

All CAS numbers were screened using the IARC database (IARC,

2021),

and we noted the classification in the Excel database (Supple-

mental File 1).

The Regulation on CLP is an EU implementation of the United Na-

tions’ Globally Harmonised System of Classification and Labelling of

Chemicals (GHS) (ECHA,

2022a, 2021).

We investigated whether the

mono-constituent substances or the multi-constituent/UVCBs, as well as

the individual constituents of the multi-constituent/UVCBs, in our

database, had harmonised carcinogenic or mutagenic classifications in

Annex VI of the CLP regulation (Carc. 1A, Carc. 1B, Carc. 2, Muta. 1A,

Muta. 1B or Muta. 2). We did this with the Annex VI Excel sheet in force

from 1 May 2020 (Annex VI to CLP_ATP13) (ECHA,

2021),

complying

with the period when the database on spray-formulated engine/brake

cleaners and lubricants was constructed. Note that in cases where we

found that substances had IARC or EU CLP classification for carcinoge-

nicity, we have not assessed whether the spray-formulated engine/brake

cleaners and lubricating agent products should be classified according to

the EU CLP criteria for mixtures (ECHA,

2019).

2.3. Danish EPA QSAR-based advisory self-classifications

QSARs are models, usually developed by machine-learning methods,

by which potential toxicities of chemical structures can be predicted.

The Danish (Q)SAR Database is a freely available online repository of

pre-calculated predictions from several free, Danish Technical

Fig. 1.

Overview of the process of evaluating the carcinogenic potential of

substances in spray-formulated engine/brake cleaners and lubricants.

BEU, Alm.del - 2021-22 - Bilag 356: Orientering om potentielt kræftfremkaldende stoffer i spraybaserede rense og smøremidler, fra beskæftigelsesministeren

J.B. Sørli et al.

Toxicology 477 (2022) 153261

University-developed and commercial QSAR models (National

Food

Institute, 2021).

For many of the included endpoints, the same training

sets are modelled in two or three QSAR systems, and so-called battery

calls based on predictions from all systems are given. Battery calls are

made on a majority vote where at least two of the three models agree

and are in the applicability domain of the models.

The Danish (Q)SAR Database has formed the basis for the Danish

EPA (Q)SAR-based advisory self-classifications of hazardous substances

(EPA_Denmark,

2022).

The applied (Q)SAR models and combinations of

these into algorithms for generic assigning of so-called advisory

self-classifications are documented in (Wedebye

et al., 2017).

The al-

gorithms contain multiple QSAR models for genotoxicity and carcino-

genicity included in the Danish (Q)SAR Database and are illustrated in

Figs. 2 and 3.

The algorithms were applied on REACH pre-registered

substances contained in the Danish (Q)SAR Database and were known

to be without any EU-harmonised classifications. It resulted in a list

published on the Danish EPA homepage with more than 54,000 sub-

stances with an advisory self-classification for at least one of the

included endpoints. Hereunder, 7323 received mutagenicity (Muta. 2),

and 4788 (of which 2023 received Muta. 2) received carcinogenicity

(Carc. 2) QSAR-based advisory self-classification. The Danish EPA

advise using the advisory self-classifications together with other reliable

information. If no other reliable information exists for a substance for

the endpoints covered by the list, the Danish EPA recommends using the

advisory self-classifications. In this project, we applied the advisory

self-classifications as is, because it was outside the scope of the project to

identify and assess possible other information. We used the Danish EPA

advisory self-classifications to identify its overlap with substances in our

database. Additionally we applied the algorithms ourselves to also

screen substances not included in the list of selected substances in the

Danish EPA project. We applied Danish EPA advisory self-classifications

to screen for possible carcinogenicity or mutagenicity potential in cases

where the substances did not have IARC or harmonised classifications.

The resulting advisory classifications that are possible with the algo-

rithms are Category 2 for mutagenicity (Muta. 2) or Category 2 for

carcinogenicity (Carc. 2). It should be noted that if a substance did not

receive an advisory self-classification for cancer or mutagenicity, it

could either be because the models predicted the substance to be

negative or because the substance was not in the defined so-called

applicability domains of the applied models. Finally, as mutagenicity

is a mechanism for carcinogenicity, we grouped mutagenic classification

with the carcinogenic classification in the QSAR-based WoE predictions.

3. Results

3.1. Database contents

Most of the products in the database were from Danish retailers’ web

stores, but several products were found in web stores in other EU and

EEA countries. Notably, nickel was found in spray-formulated

Fig. 3.

Schematic diagram illustrating the QSAR models and algorithm applied

to assign Danish EPA advisory classifications for carcinogenicity.

lubricants, but only from web stores in the USA and Australia. Thus,

these products were not included.

We identified 123 spray-formulated engine/brake cleaners and

lubricating products available from EU websites. Of these, 82 had Safety

Data Sheets (SDSs) readily available on the internet. The remaining

products were discarded, based on the lack of direct access on the

internet. In total, we identified 376 different CAS numbers; 88 were

substances directly listed in the CAS number list in the SDS (14 inorganic

elements and 74 organic molecules) (Fig.

Supplemental File 1). Forty-

four multi-constituent/ UVCBs were identified, mostly including hy-

drocarbons (mineral oils). When these multi-constituent/UVCBs were

unravelled for constituent CAS numbers using the ECHA registration

dossiers, we identified 244 additional substances (illustrated in

Fig. 4).

The inorganic substances are listed in

Table 1.

3.2. Overview of the carcinogenic potential of the substances

As shown in

Fig. 5

(upper panel), only a few substances are classified

as IARC 1 or 2B (we identified no Group 2A substances). One substance,

diethanolamine (Group 2B), was listed directly in the SDS in one prod-

uct, while two substances were IARC 1 and part of multi-constituent/

UVCBs and three substances were IARC 2B and part of multi-

constituent/UVCBs (Fig.

substance identities provided in

Table 2).

EU-harmonised classification identified two substances that are Carc.

2 and listed directly as ingredients, while 10 were Carc. 1A, 1B or 2 and

Fig. 2.

Schematic diagram illustrating the QSAR models and algorithm applied

to assign Danish EPA advisory self-classifications for mutagenicity. *The

training set data were not used for the cancer and chromosomal aberrations in

CHO cells models as these were proprietary information in the commer-

cial models.

Fig. 4.

Overview of the type of substances.

BEU, Alm.del - 2021-22 - Bilag 356: Orientering om potentielt kræftfremkaldende stoffer i spraybaserede rense og smøremidler, fra beskæftigelsesministeren

J.B. Sørli et al.

Toxicology 477 (2022) 153261

Table 1

Metals and other inorganic elements in the products *Nickel was identified in

two products from outside the EU and thus not included in the current database.

Number of substances that were

metals or other inorganic

elements*

14 of 376

substances (3%)

Aluminium

Ammonia

Boron (nitride)

Calcium dihydroxide

Copper

Lithium

Magnesium (silicate)

Molybdenum

(disulphide)

Phosphorodithioic

acid

Potassium

(hydroxide)

Sulphur

Tungsten

Zinc (sulphide)

Zinc salts

n/a

Number of products containing

metals or other elements

18 of 82 products

(22%)

substances in a multi-constituent/UVCB (Fig.

5).

Notably, in the CLP

regulation, there is also harmonised classification of the multi-

constituent/UVCBs themselves: 5 multi-constituent/UVCBs had a

harmonised classification Carc. 1A and 17 had a Carc. 1B, while 22

multi-constituent/UVCBs had no harmonised classification (Fig.

5).

total, 28 of 44 multi-constituent/UVCBs had a carcinogenic potential

based on the 22 with direct harmonised classification and based on

constituents with potential by either IARC, harmonised classification or

QSAR. For the multi-constituent/UVCBs that themselves had either

Carc. 1A or Carc. 1B, many were present in products at a content of over

50% (numbers given for each multi-constituent/UVCB in

Table 2).

When we searched for the substances within our database in the

Danish EPA QSAR-based advisory list for the self-classification of

chemical substances (Fig.

lower panel), two substances had a QSAR-

based Danish EPA advisory self-classification mutagenic or carcinogenic

and were listed directly in the SDS. Seven substances that fulfilled the

QSAR-based advisory self-classifications were part of multi-constituent/

UVCBs. In addition, 229 substances were in the Danish (Q)SAR Data-

base, but without a QSAR-based advisory self-classification for Muta.2

or Carc.2, and this could either be due to the QSAR models not pre-

dicting the substances to be (sufficiently) positive, or the substances

could be outside the so-called applicability domains of the models.

Moreover, 94 substances were not present in the Danish (Q)SAR Data-

base. Those not found in the Danish (Q)SAR database included: 1) multi-

constituent/UVCBs, 2) substances too small to be eligible for QSAR, 3)

metals or other inorganic substances (Table

1).

The database only con-

tains organic molecules with two carbons or more (and only certain

other atoms). Details of the QSAR-prioritised substances can be found in

Table 2.

One substance, 2-(1-methoxypropan-2-yloxy)propan-1-ol, was

not pre-registered in REACH and thus not present in the screening list for

the Danish EPA QSAR-based advisory list for self-classification of

chemical substances (ECHA,

2022b).

We manually ran the QSAR algo-

rithms for advisory self-classification for this substance and found it to

be Muta.2 (Table

2).

Figs. 6 and 7

show how much additional information was gained by

supplementing IARC evaluations with harmonised classifications (9

substances) and QSAR predictions (9 substances) at the mono-

constituent level (Fig.

6).

At the multi-constituent/UVCB level (Fig.

7),

11 had a carcinogenic potential based on IARC, 16 additional ones based

on harmonised classification, and one by QSAR (Fig.

7).

3.3. Overview of the carcinogenic potential at the product level

At the product level,

Fig. 8

illustrates the number of products in

Fig. 5.

The carcinogenic potential of ingredient substances as determined by

IARC classification, EU-harmonised classification and QSAR-based advisory

self-classification. Concerning harmonised classifications, none of the sub-

stances in our database were classified for mutagenicity, while not also classi-

fied for carcinogenicity. Thus, the mutagenicity classification is omitted from

the pie chart.

different categories of concern. The first group comprises 31 products

with either IARC 1 or 2B classified substances and seven products with

harmonised classified Carc. substances that were not already IARC

classified (IARC class 1 or 2B). One additional product has no IARC 1 or

2B substances or harmonised classified substances, but concern can be

expressed for this product based on QSAR. A few products contain in-

gredients with no CAS number or for which the composition of multi-

constituent/UVCBs cannot be unravelled because of limited informa-

tion in ECHA registration dossiers (Fig.

8).

Finally, there are products for

which the composition is provided, and their reported composition does

not contain substances that IARC classifies, have harmonised classifi-

cation, or raise a concern based on QSAR. In the lower panels of

Fig. 8,

the results are divided into the two product categories 1) Engine/brake

BEU, Alm.del - 2021-22 - Bilag 356: Orientering om potentielt kræftfremkaldende stoffer i spraybaserede rense og smøremidler, fra beskæftigelsesministeren

J.B. Sørli et al.

Toxicology 477 (2022) 153261

Table 2

Overview of the substances evaluated to have carcinogenic potential.

Classification

IARC classifications

Substances identified as IARC Group

1 (two substances)

Substance name/CAS number

Constituents in multi-constituent/UVCB

Benzene / 71–43–2

Buta-1,3-diene / (1,3 butadiene) /

106–99–0

–

Substances listed directly in SDS

Diethanolamine / 111–42–2 (concentration

in one product 1–3%)

Constituents in a multi-constituent/UVCB

Ethylbenzene / 100–41–4

Cumene / 98–82–8

Naphtalene / 91–20–3

Substances listed directly in SDS

Toluene / 108–88–3

1,4-Dihydroxybenzene/ 123–31–9

Butylglycol / 2-butoxyethanol / 111–76–2

Xylene / 1330–20–7

Coumarin/91–64–5

Constituents in a multi-constituent/UVCB

Propene / 115–07–1

Toluene / 108–88–3

Xylenes / various CAS numbers

Pyrene / 129–00–0

Substances listed directly in SDS

None

Constituents in a multi-constituent/UVCB

Benzene71–43–2

Buta-1,3-diene / (1,3 butadiene) /

106–99–0

Multi-constituent/UVCBs

Lubricating oils 74869–22–0 (5–15% in one

product)

Hydrocarbons, C3–4-rich, petroleum

distillate 68512–91–4 (50–75% in two

products)

Hydrocarbons, C3–4 68476–40–4 (30–50%

in two products)

Hydrocarbons, C4 87741–01–3 (only listed

as a constituent of one UVCB, 68476–85–7,

in one product)

Petroleum gases, liquefied 68476–85–7 (in

four products at up to 30–60%)

Substances listed directly in SDS

None

Constituents in a multi-constituent/UVCB

Benzo[j]fluoranthene 205–82–3

Benzo[e]acephenanthrylene 205–99–2

Benzo[k]fluoranthene 207–08–9

Chrysene 218–01–9

Benzo[def]chrysene 50–32–8

Dibenz[a,h]anthracene 53–70–3

Benz[a]anthracene 56–55–3

Multi-constituent/UVCBs

Petrolatum8009–03–8 (only listed as a

constituent of one UVCB, 8042–47–5, in

one product)

Hydrocarbons, C6-C7, n-alkanes,

isoalkanes, cyclics,

5% n-hexane

64742–49–0 (in 15 products at up to

50–100%)

Distillates (petroleum), solvent-refined

heavy paraffinic 64741–88–4 (1 product,

per centage not reported)

Distillates Petroleum Hydrotreated

64742–46–7 (1 product, per centage not

reported)

Hydrocarbons, C9-C11, n-alkanes,

isoalkanes, cyclics,

2% aromatics

64742–48–9 (11 products, up to 50–100%)

Distillates (petroleum), hydrotreated heavy

naphthenic 64742–52–5 (1 product, per

Table 2

(continued )

centage not reported)

Petroleum Base Oil 64742–53–6 (1 product,

per centage not reported)

Distillates (petroleum), hydrotreated heavy

paraffinic 64742–54–7 (2 products, per

centage not reported)

Petroleum Base Oil 64742–56–9 (one

product, per centage not reported)

Distillates (petroleum), solvent-dewaxed

heavy paraffinic 64742–65–0

Petroleum Base Oil 64742–71–8 (one

product, per centage not reported)

Naphtha (petroleum), hydrodesulfurised

heavy 64742–82–1 (2 products at 1–1.5%

and 40–50%

Solvent-nafta (petroleum), slightly

aromatic (benzene

<0.1%)

64742–95–6

(two products 10–25% in both)

Alkanes, C12–26-branched and linear

90622–53–0 (One product at 40–50%)

Aromatic hydrocarbons, C890989–38–1

(only listed as a constituent of one UVCB,

64742–95–6, in one product)

Naphtha (petroleum), hydrodesulfurised

light, dearomatised 92045–53–9 (1 product

at 20–50%)

Distillates (petroleum), hydrotreated light

paraffinic 64742–55–8 (1 product at

25–50%)

Substances identified as Carc. 2 (three

Substances listed directly in SDS

mono-constituent substances)

1,4-Dihydroxybenzene 123–31–9 (<0.1%

in one product)

Dichloromethane 75–09–2 (50–70% in one

product)

Constituents in a multi-constituent/UVCB

Naphthalene91–20–3

Substances identified as Muta. 1B

Constituent in a multi-constituent/UVCB

(one mono-constituent substance

Buta-1,3-diene106–99–0

and 11 multi-constituent/UVCBs)

Multi-constituent/UVCBs

Hydrocarbons, C9-C11, n-alkanes,

isoalkanes, cycloalkanes 64742–48–9

Hydrocarbons, C7, n-alkanes, isoalkanes,

cyclics 64742–49–0

Naphtha (petroleum), hydrodesulfurised

heavy 64742–82–1

Solvent-nafta (petroleum), slightly

aromatic (benzene

<0.1%)

64742–95–6

Petroleum gases, liquefied 68476–85–7

Hydrocarbons, C3–4-rich, petroleum

distillate 68512–91–4

Naphtha (petroleum), hydrodesulfurised

light, dearomatised 92045–53–9

Benzo[def]chrysene 50–32–8

Hydrocarbons, C3–468476–40–4

Hydrocarbons, C4 87741–01–3

Aromatic hydrocarbons, C890989–38–1

Substances identified as Muta. 2 (two

Substances listed directly in SDS

mono-constituent substances)

1,4-Dihydroxybenzene 123–31–9 (<0.1%

in one product)

Constituents in a multi-constituent/UVCB

Chrysene 218–01–9

QSAR-based advisory CLP classifications

Substances identified as Carc. 2

Substances listed directly in SDS

advisory classification (two mono-

Coumarin / 2 H-1-benzopyran-2-one /

constituent substances)

91–64–5 (<1% in one product)

Constituents in a multi-constituent/UVCB

Benzo[ghi]perylene / 191–24–2

Substances with Muta. 2 advisory

Substances listed directly in SDS

classification (seven mono-

Dimethoxymethan / 109–87–5 (80–95% in

constituent substances)

one product)

Constituents in a multi-constituent/UVCB

Fluoranthene 206–44–0

Fluorene / 86–73–7

2-(1-methoxypropan-2-yloxy)propan-1-ol /

55956–21–3

Indeno[1,2,3-cd]pyrene / 193–39–5

(continued

on next page)

Substances identified as IARC Group

2A (none)

Substances identified as IARC Group

2B (four mono-constituent

substances)

Substances identified as IARC Group

3 (seven mono-constituent

substances)

EU-harmonised classifications

Substances identified as Carc. 1A (two

mono-constituent substances and

five multi-constituent/UVCBs)

Substances identified as Carc. 1B

(seven mono-constituent

substances and 16 multi-

constituent/UVCBs)

BEU, Alm.del - 2021-22 - Bilag 356: Orientering om potentielt kræftfremkaldende stoffer i spraybaserede rense og smøremidler, fra beskæftigelsesministeren

J.B. Sørli et al.

Toxicology 477 (2022) 153261

Table 2

(continued )

Pyrene / 129–00–0

4-vinyl-m-xylene / 2234–20–0

Notably, ethanol 64–17–5 was identified in the database and is classified

Group 1 by IARC as an ingredient in alcoholic beverages. Yet we have not

included it in the table because of the inhalation route being considered here.

Fig. 6.

Additional information on mono-constituent substances was obtained

by supplementing IARC with harmonised classification and QSAR. The overlap

between the different classifications illustrated how many additional substances

were identified by starting with IARC and then a) adding harmonised classifi-

cation and b) adding QSAR classification.

Fig. 7.

The carcinogenic potential in multi-constituent/UVCBs: Additional in-

formation was obtained by supplementing IARC with harmonised classification

and QSAR. The overlap between the different classifications illustrated how

many additional substances were identified by starting with IARC and then a)

adding harmonised classification and b) adding QSAR classification.

cleaners and 2) lubricants. For cleaners, 35% (17/48) have carcinogenic

potential, with lubricants having a larger proportion, 56% (25/44).

4. Discussion

4.1. Fraction of the investigated substances with a carcinogenic potential

IARC classifications 1 and 2B were found for six substances.

Harmonised classification pointed to 12 substances that were Carc. 1A,

1B or Carc. 2 and 22 multi-constituent/UVCBs (1A or 1B based on CAS

numbers of these multi-constituent/UVCBs). QSAR pointed to nine

substances with carcinogenic potential. Taking the overlap in substances

identified by the three screening methods into account, this gives a total

24 of 376 substances that had a carcinogenic potential. Thus, the frac-

tion of mono-constituent substances with a carcinogenic potential

(compared with the total number of substances) is approximately 6%. In

addition, we have another 6% when we look at multi-constituent/

Fig. 8.

Illustration of the number of products potentially containing muta-

genic/carcinogenic substances. There may be substances of concern in the pink

box (products with multi-constituent/UVCBs of unknown composition) as we

were not able to unravel the composition of these CAS numbers. The two lower

panels represent product groups. Notably, some products are counted twice as

they are marketed as so-called multipurpose products being both cleaners

and lubricants.

UVCBs based on EU-harmonised classification. Overall, if we look at

the substance level, we find a smaller fraction with a carcinogenic po-

tential than compared to the situation on the product level

–

this will be

discussed below.

When we look at the substances with carcinogenic potential, the

potential is mainly found in multi-constituents/UVCBs. Only five sub-

stances with carcinogenic potential are listed directly in the SDS

(including one with Muta. 2 based on QSAR). At the same time, the other

19 are part of multi-constituent/UVCBs (22 multi-constituent/UVCBs

are placed in Carc.-groups based on EU-harmonised classification).

This finding suggests that most of the carcinogenic potential is found in

multi-constituent/UVCBs, and that single substances with a carcino-

genic potential have already been eliminated from products by historic

BEU, Alm.del - 2021-22 - Bilag 356: Orientering om potentielt kræftfremkaldende stoffer i spraybaserede rense og smøremidler, fra beskæftigelsesministeren

J.B. Sørli et al.

Toxicology 477 (2022) 153261

safe-by design decisions. For instance, we did not find trichloroethylene

(an IARC 1 substance). It has a

“Sunset

date” of 21 April 2016 for

authorised use (EU,

2013)

and 1 March 2023 for the use of the substance

in the production of spare parts as articles or as complex products for the

repair of articles or complex products (ECHA,

2016).

In addition, in

recent years, changes have been made in the workplace to limit the use

of solvents. For example, in Japan, the use of aromatic solvents fell over

time (Ukai

et al., 2014).

Moreover, we note that nickel, classified IARC

2B, was only identified in spray products from outside the EU and thus

not included in the database. Finally, the carcinogenic substance

diethanolamine was found only in one product in our database, sug-

gesting a very limited use.

The fact that multi-constituent/UVCBs hold a carcinogenic potential

and are still used could reflect that it is difficult to obtain products that

are effective cleaners (degreasers) and lubricating agents without these

substances. Nonetheless, the question is whether one can choose the

more refined mineral oils in the products and thus have lower carcino-

genic potential. This point will be discussed further below, and is rele-

vant taking into account that at the product level there is a substantial

carcinogenic potential.

4.2. Carcinogenic potential at the product level

At the product level, a substantially larger fraction (48%) is poten-

tially carcinogenic, as compared to the fraction seen at the substance

level (6%) (plus 6% of the CAS numbers in the form of multi-

constituent/UVCBs, based on EU-harmonised classification). This

finding reflects that the substances with a carcinogenic potential are

more frequently used in the products than those with no identified

carcinogenic potential. It likely reflects the notion that the most bio-

logically active substances are also those most efficient in cleaning

(degreasing) or lubrication. On the product level, the extent of sub-

stances with IARC classification, harmonised classification or QSAR

predicted carcinogenic potential is higher for lubricating agents (58%)

than it is for cleaners (35%) (Fig.

8).

Notably, it is difficult to determine whether products with no iden-

tified carcinogenic potential indeed do not have such a potential or

whether it has just not been detected. One challenge is the multi-

constituent/UVCBs with no composition given in ECHA registration

dossiers or with generic names with no CAS number provided. We also

found that for some products, the main substances included in the name

of the products were not provided as substances in the list of substances

in the SDS

–

as was seen for several metals. Yet another challenge is that

for a substantial number of products, i.e., 41 of 123 in the database, we

could not readily retrieve an SDS on the internet. It is unclear if products

sold by retailers who do not provide an SDS readily downloadable on the

internet are marketing better or worse products in terms of carcinogenic

potential. Also, we note that we did not look at whether the SDSs were in

accordance with the regulation. For substances that had a carcinogenic

potential, we returned to the SDSs and retrieved the concentrations in

the products. For the 22 multi-constituent/UVCBs that themselves had

either Carc. 1A or Carc. 1B harmonised classification, many were pre-

sent in products at over 50% (numbers given for each multi-constituent/

UVCB in

Table 2).

For the mono-constituent substances the concentra-

tions were: diethanolamine (1–3%, 1 product), dichloromethane

(50–70% 1 product), dimethoxymethan (80–95%, 1 product), 1,4-dihy-

droxybenzene (<0.1%, 1 product), and coumarin (<1%, 1 product).

Overall, these numbers suggest that some products present ingredients

with a carcinogenic potential in substantial amounts. This point is

particularly seen for multi-constituents/UVCBs present in many

products.

4.3. Reducing exposure to products with carcinogenic potential

4.3.1. Are there any hydrocarbon multi-constituent/UVCBs that do not

contain substances with a carcinogenic potential?

Hydrocarbon mixtures are commonly included in many products.

From a safe-by-design point of view, it would be desirable if we could

state that one or more hydrocarbon mixtures were more safe than others

to include in the products. Few hydrocarbon UVCBs did not contain any

substances with carcinogenic potential, while many hydrocarbon UVCBs

indeed had a carcinogenic potential. Based on this, it is questionable

whether one can justify using specific UVCBs instead of others in a safe-

by design approach

–

without further research. Nonetheless, one aspect

reviewed by IARC is that some mineral oils that were unrefined or only

mildly refined were IARC Group 1 (IARC,

2012),

while more refined

ones were IARC Group 3 (IARC,

1984).

Notably, this statement dates

back to 1984. Yet overall, based on considerations by IARC and the

notion that refinement generally removes unwanted substances, one

safe-by design measure to limit the carcinogenic potential of these

products could be to consider more-refined hydrocarbon mixtures.

4.3.2. Avoiding future substances that hold a carcinogenic potential - the

contribution of QSAR

While IARC and harmonised classification look into the already-

existing human knowledge and experimental assays in animals and

cells, QSAR infers the potential toxicity based on the chemical structure.

QSAR can thus be used to assess the carcinogenic potential of sub-

stances, already before synthesis, and on many structures at a low cost.

Thus, QSAR provides an opportunity to predict potential future long-

term effects.

In the current overview looking at the substance level, the QSAR

WoE algorithm identified nine substances that were not already classi-

fied as IARC 1 or 2 or had a harmonised classification as carcinogenic or

mutagenic. Overall, QSAR contributed substantially to screening the

carcinogenic potential of the substances in products. Although when

looking at the product level, there was a substantial overlap between

IARC/harmonised classification and QSAR predictions (Table

2),

reflecting that other substances in the products already had an IARC or

EU-harmonised classification. Nevertheless, the situation may be

different when using this setup for product groups other than engine

maintenance. For instance, we have previously demonstrated the value

of QSAR in prioritising potential asthma-inducing substances (Hadrup

et al., 2022a).

4.3.3. Avoiding exposure by not using spray-formulation

Besides controlling substances with a carcinogenic potential in the

products, another option is to limit exposure by not formulating the

products in spray form. This step will prevent aerosols from being

inhaled. Some of the studied products indeed had alternative formula-

tions from the retailer that could be applied in other ways, e.g., a paste.

Choosing these products would already be a possibility for workers

today to reduce exposure by inhalation, although with the caveat that

gloves would be needed to avoid skin exposure. Yet one thing that we

encountered through personal communication with mechanics is the

widespread use of compressed air for cleaning engine parts

–

something

that aerosolises surface-deposited grease and other substances, once

again providing potential for occupational lung exposure.

4.4. Comparison with previous studies on exposure to engine/brake

cleaners and lubricants

One review study from 2000 collected exposure information on some

17,000 hydrocarbon solvent exposure measurements in similar end-use

products (painting and coating, printing, and adhesives). The authors

found that reported hydrocarbon solvent exposures fell four-fold from

1960 to 1998 (Caldwell

et al., 2000).

We found no other investigations

describing the exposure and risk characterisation in engine/brake

BEU, Alm.del - 2021-22 - Bilag 356: Orientering om potentielt kræftfremkaldende stoffer i spraybaserede rense og smøremidler, fra beskæftigelsesministeren

J.B. Sørli et al.

Toxicology 477 (2022) 153261

cleaners and lubricants. This point highlights the importance of the

current work on carcinogenic hazard characterisation of

spray-formulated engine/brake cleaners and lubricants.

4.5. Knowledge gaps, research recommendations and limitations of the

study

In the current work, we identified a set of substances in spray-

formulated engine/brake cleaners and lubricants with carcinogenic

potential. A next step could be the collection of further data on their

carcinogenic effect, e.g. whether they act by a threshold or non-

threshold genotoxic mechanism, which again informs the risk assess-

ment process. Further research into whether more refined hydrocarbon

mixtures have lower carcinogenic potential would be useful to inform

safe-by-design processes. Another research area is substances only

identified by the QSAR algorithm, which need further in-depth study

(literature search, IATA WoE, experimental testing and/or read-across)

to assess whether they are carcinogenic. In addition, it is important to

consider potential additive or even synergistic mixture effects of

different ingredients in these products, which often contain numerous

substances (Hadrup,

2014; Hadrup et al., 2016, 2015; Kortenkamp,

2007; Olmstead and LeBlanc, 2005).

Finally, inorganic elements are not

covered by QSAR, and for those readings, we rely on IARC evaluations

and the toxicological literature. As an example of the latter, we recently

published literature reviews of four of the metals present in the products

in the current database, boron nitride, tungsten, molybdenum and

lithium (Hadrup

et al., 2022b, 2021).

5. Conclusion

This work provides an overview of the extent of genotoxic and

carcinogenic substances in degreasers and lubricating spray-formulated

products. We conclude that in the investigated EU marketed spray-

formulated engine/brake cleaners and lubricants, 24 of 332 mono-

constituent substances and 28 of 44 multi-constituent/UVCBs had a

carcinogenic potential based on IARC, EU-harmonised classification,

and QSAR. At the product level, 39 of 82 contained substances with an

identified carcinogenic potential by either IARC, harmonised classifi-

cation or QSAR, suggesting a carcinogenic potential in half of the

products for which sufficient information on composition was available.

These results can contribute to safe-by-design choices for manufacturers.

Also, the data informs that precautions should be taken at the workplace

to limit exposure to these agents. One possibility is to choose alternative

application methods to spraying.

Declaration of Competing Interest

The authors declare the following financial interests/personal re-

lationships which may be considered as potential competing interests:

Niels Hadrup reports financial support was provided by The Danish

Working Environment Research Fund.

Acknowledgements

This work was financed by a grant from The Danish Working Envi-

ronment Research Fund (project name Sikker-Motor; grant number:

29–2019-09).

CRediT authorship contribution statement

Jorid B. Sørli:

Conceptualization, Data curation, Writing

–

review

editing, Funding acquisition.

Marie Frederiksen:

Conceptualization,

Data curation, Writing

–

review

editing, Funding acquisition.

Nikolai

G. Nikolov:

Data curation, Methodology, Software, Writing

–

review

editing.

Eva B. Wedebye:

Data curation, Methodology, Software,

Writing

–

review

editing.

Niels Hadrup:

Conceptualization, Formal

analysis, Investigation, Data curation, Writing

–

original draft, Writing

–

review

editing, Visualization, Project administration, Funding

acquisition.

Appendix A. Supporting information

Supplementary data associated with this article can be found in the

online version at

doi:10.1016/j.tox.2022.153261.

References

Caldwell, D.J., Armstrong, T.W., Barone, N.J., Suder, J.A., Evans, M.J., 2000.

Hydrocarbon solvent exposure data: compilation and analysis of the literature.

AIHAJ - Am. Ind. Hyg. Assoc. 61, 881–894.

https://doi.org/10.1080/

15298660008984602.

ECHA, 2022a. Classification and labelling (CLP/GHS) [WWW Document]. URL https://

ec.europa.eu/growth/sectors/chemicals/classification-and-labelling-clpghs_da.

ECHA, 2022b. REACH pre-registered substances [WWW Document]. URL https://echa.

europa.eu/da/information-on-chemicals/pre-registered-substances?p_p_id=disslists_

WAR_disslistsportlet&p_p_lifecycle=1&p_p_state=normal&p_p_mode=view&_

disslists_WAR_disslistsportlet_javax.portlet.action=searchDissLists.

ECHA, 2021. ECHA CLP [WWW Document]. URL https://echa.europa.eu/da/

information-on-chemicals/annex-vi-to-clp.

ECHA, 2019. Introductory Guidance on the CLP Regulation.

ECHA, 2016. Trichlorethylene [WWW Document]. URL https://echa.europa.eu/

authorisation-list/-/dislist/details/0b0236e1807e1a4f.

EPA_Denmark, 2022. Danish EPA [WWW Document]. URL Danish EPA homepage,

Chemicals/Chemicals in products/Assessment of Chemicals/The advisory list for

self-classification https://eng.mst.dk/chemicals/chemicals-in-products/assessment-

of-chemicals/the-advisory-list-for-self-classification-of-hazardous-substa (accessed

5.27.22).

EU, 2013. COMMISSION REGULATION (EU) No 348/2013.

Hadrup, N., 2014. Evidence from pharmacology and pathophysiology suggests that

chemicals with dissimilar mechanisms of action could be of bigger concern in the

toxicological risk assessment of chemical mixtures than chemicals with a similar

mechanism of action. Regul. Toxicol. Pharmacol. 69, 281–283.

https://doi.org/

10.1016/j.yrtph.2014.05.007.

Hadrup, N., Frederiksen, M., Sharma, A.K., 2021. Toxicity of boric acid, borax and other

boron containing compounds: a review. Regul. Toxicol. Pharmacol. 121, 104873

https://doi.org/10.1016/j.yrtph.2021.104873.

Hadrup, N., Frederiksen, M., Wedebye, E.B., Nikolov, N.G., Carøe, T.K., Sørli, J.B.,

Frydendall, K.B., Liguori, B., Sejbaek, C.S., Wolkoff, P., Flachs, E.M., Schlünssen, V.,

Meyer, H.W., Clausen, P.A., Hougaard, K.S., 2022a. Asthma-inducing potential of 28

substances in spray cleaning products-Assessed by quantitative structure activity

relationship (QSAR). Test. Lit. Rev. J. Appl. Toxicol. 42, 130–153.

https://doi.org/

10.1002/jat.4215.

Hadrup, N., Pedersen, M., Skov, K., Hansen, N.L., Berthelsen, L.O., Kongsbak, K.,

Boberg, J., Dybdahl, M., Hass, U., Frandsen, H., Vinggaard, A.M., 2015.

Perfluorononanoic acid in combination with 14 chemicals exerts low-dose mixture

effects in rats. Arch. Toxicol.

https://doi.org/10.1007/s00204-015-1452-6.

Hadrup, N., Sørli, J.B., Sharma, A.K., 2022b. Pulmonary toxicity, genotoxicity, and

carcinogenicity evaluation of molybdenum, lithium, and tungsten: A review.

Toxicology 467, 153098.

https://doi.org/10.1016/j.tox.2022.153098.

Hadrup, N., Svingen, T., Mandrup, K., Skov, K., Pedersen, M., Frederiksen, H.,

Frandsen, H.L., Vinggaard, A.M., 2016. Juvenile male rats exposed to a low-dose

mixture of twenty-seven environmental chemicals display adverse health effects.

PLoS One 11, e0162027.

https://doi.org/10.1371/journal.pone.0162027.

IARC, 2021. List of Classifications Agents classified by the IARC Monographs, Volumes

1–128 [WWW Document]. URL https://monographs.iarc.who.int/list-of-

classifications/.

IARC, 2014. Trichloroethylene, Tetrachloroethylene, and Some Other Chlorinated

Agents. Volume 106.

IARC, 2012. Chemical Agents and Related Occupations IARC Monographs on the

Evaluation of Carcinogenic Risks to Humans Volume 100F.

IARC, 1984. IARC Monographs on the evaluation of the carcinogenic risk of chemicals to

humans, Volume 33.

Kaerlev, L., Hansen, J., Hansen, H.L., Nielsen, P.S., 2005. Cancer incidence among

Danish seafarers: a population based cohort study. Occup. Environ. Med. 62,

761–765.

https://doi.org/10.1136/oem.2005.020818.

Kortenkamp, A., 2007. Ten years of mixing cocktails: a review of combination effects of

endocrine-disrupting chemicals. Environ. Health Perspect. 115, 98–105.

https://doi.

org/10.1289/ehp.9357.

National_Food_Institute, T., 2021. Danish (Q)SAR Database [WWW Document]. URL

http://qsar.food.dtu.dk/ (accessed 4.9.21).

Olmstead, A.W., LeBlanc, G.A., 2005. Toxicity assessment of environmentally relevant

pollutant mixtures using a heuristic model. Integr. Environ. Assess. Manag. 1,

114–122.

Pukkala, E., Martinsen, J.I., Lynge, E., Gunnarsdottir, H.K., Spar

n, P., Tryggvadottir, L.,

Weiderpass, E., Kjaerheim, K., 2009. Occupation and cancer - follow-up of 15 million

people in five Nordic countries. Acta Oncol. 48, 646–790.

https://doi.org/10.1080/

02841860902913546.

BEU, Alm.del - 2021-22 - Bilag 356: Orientering om potentielt kræftfremkaldende stoffer i spraybaserede rense og smøremidler, fra beskæftigelsesministeren

J.B. Sørli et al.

Ugelvig Petersen, K., Volk, J., Kaerlev, L., Lyngbeck Hansen, H., Hansen, J., 2018. Cancer

incidence among merchant seafarers: an extended follow-up of a Danish cohort.

Occup. Environ. Med. 75, 582–585.

https://doi.org/10.1136/oemed-2018-105037.

Toxicology 477 (2022) 153261

Ukai, H., Samoto, H., Nagasawa, Y., Ohashi, F., Ikeda, M., 2014. Shift to non-aromatic

solvents in solvent workplaces in Japan. Environ. Health Prev. Med. 19, 245–248.

https://doi.org/10.1007/s12199-014-0383-9.

Wedebye, E., Nikolov, N., Løfsted, M., Tyle, H., 2017. Brief documentation on the (Q)SAR

models and the algorithms used for the advisory list for self-classification.