formaldehyd2.pdf

Contact Dermatitis 2010: 62: 18–31

CONTACT DERMATITIS

Review Article

Formaldehyde-releasers in cosmetics: relationship

to formaldehyde contact allergy

Part 2. Patch test relationship to formaldehyde contact allergy,

experimental provocation tests, amount of formaldehyde released,

and assessment of risk to consumers allergic to formaldehyde

ANTON DE GROOT 1 , IAN R. WHITE 2 , MARI-ANN FLYVHOLM 3 , GERDA LENSEN 1 AND PIETER-JAN COENRAADS 1

1 Department of Dermatology, University Medical Center Groningen, University of Groningen,

The Netherlands,

2 Department of Cutaneous Allergy, St John’s Institute of Dermatology, St. Thomas’ Hospital, London, UK, and

3 National Research Centre for the Working Environment, Copenhagen, Denmark

This is the second part of an article on formaldehyde-releasers in cosmetics. The patch test relationship

between the releasers in cosmetics to formaldehyde contact allergy is reviewed and it is assessed

whether products preserved with formaldehyde-releasers may contain enough free formaldehyde to pose

a threat to individuals with contact allergy to formaldehyde. There is a clear relationship between

positive patch test reactions to formaldehyde-releasers and formaldehyde contact allergy: 15% of

all reactions to 2-bromo-2-nitropropane-1,3-diol and 40 – 60% of the reactions to the other releasers

are caused by a reaction to the formaldehyde in the test material. There is only fragmented data

on the amount of free formaldehyde in cosmetics preserved with formaldehyde donors. However,

all releasers (with the exception of 2-bromo-2-nitropropane-1,3-diol, for which adequate data are

lacking) can, in the right circumstances of concentration and product composition, release > 200 p.p.m.

formaldehyde, which may result in allergic contact dermatitis. Whether this is actually the case in

any particular product cannot be determined from the ingredient labelling. Therefore, we recommend

advising patients allergic to formaldehyde to avoid leave-on cosmetics preserved with quaternium-15,

diazolidinyl urea, DMDM hydantoin, or imidazolidinyl urea, acknowledging that many would tolerate

some products.

Key words: benzylhemiformal; 5-bromo-5-nitro-1,3-dioxane; 2-bromo-2-nitropropane-1,3-diol;

cosmetics; diazolidinyl urea; DMDM hydantoin; formaldehyde; imidazolidinyl urea; preservative;

quaternium-15; sodium hydroxymethylglycinate.

John Wiley & Sons A/S, 2010.

Accepted for publication 22 July 2009

In the ﬁrst part of this article (1), key data on

formaldehyde-releasers used in cosmetics (benzyl-

hemiformal, 5-bromo-5-nitro-1,3-dioxane, 2-bromo-

2-nitropropane-1,3-diol, diazolidinyl urea, DMDM

hydantoin, imidazolidinyl urea, quaternium-15 and

sodium hydroxymethylglycinate) were presented,

including applications, frequency of sensitization,

relevance of patch test reactions, and their frequency

of use in cosmetics. In this second part , the patch

test relationship of these chemicals to formalde-

hyde are reviewed and assessed whether products

preserved with formaldehyde-releasers may con-

tain enough free formaldehyde to pose a threat to

individuals who are contact allergic to formaldehyde.

Contact Dermatitis 2010: 62: 18–31

FORMALDEHYDE-RELEASERS IN COSMETICS: PART 2 19

Relationship between Formaldehyde-Releasers

and Contact Allergy to Formaldehyde

When patients are tested with formaldehyde and

one or more formaldehyde-releasers, concomitant

positive reactions may be observed to formalde-

hyde and (one or more) donors. It is often assumed

that this is due to ‘cross-allergy’ to formalde-

hyde (more correctly termed ‘pseudo-cross-allergy’

in our opinion), i.e. that the patch test reac-

tion to the donor is caused by the formalde-

hyde present in the test material of the releaser.

In favour of this assumption is that often more

than one releaser reacts in formaldehyde-sensitive

patients, even when the preservatives are not struc-

turally related (cross-reactions unlikely). In a recent

study, for example, among 219 patients aller-

gic to formaldehyde and/or formaldehyde-releasers

(quaternium-15, diazolidinyl urea, imidazolidinyl

urea, 2-bromo-2-nitropropane-1,3-diol, DMDM

hydantoin, and ethylene urea/melamine-formaldehy-

de resin), 76 (35%) were positive to only one of

these compounds, 53 (24%) reacted to two, 27

(12%) reacted to three, 31 (14%) to four, 25 (11%)

to ﬁve, and 7 (3%) patients reacted to six of the

tested formaldehyde(releasing) materials (2). Of 59

patients, patch test positive to quaternium-15 1%

petrolatum (pet.), 31 (53%) were also allergic to

formaldehyde (tested 2% aqua). However, of seven

patients allergic to quaternium AND to 2-bromo-2-

nitropropane-1,3-diol, imidazolidinyl urea, or both,

ALL were also allergic to formaldehyde, attest-

ing to the role of formaldehyde in multiple reac-

tions to releasers (3). In a USA study with 454

patients allergic to formaldehyde, 38 patients (8.4%)

reacted to both formaldehyde and two releasers

and 4 reacted to all three donors tested (2-bromo-

2-nitropropane-1,3-diol, imidazolidinyl urea, and

quaternium-15). Of these patients reacting to two

or three releasers (suggesting formaldehyde to be

the cause), quaternium-15 (which probably releases

the most formaldehyde) was positive in ALL (4).

Perret and Happle investigated 13 patients with pos-

itive patch test reactions to diazolidinyl urea 2%

aqua. Six of them were also allergic to formalde-

hyde. All 13 patients were additionally tested with

the formaldehyde-releasers urea formaldehyde (8%

aqua), glyoxal urea (10% aqua), and quaternium-15

(1% pet.). All six patients reacting to diazolidinyl

urea AND to formaldehyde also reacted to the

other releasers, whereas of those patients not aller-

gic to formaldehyde, NONE had a positive reaction

to any of the formaldehyde-releasers. It was con-

cluded that in the six patients allergic to diazolidinyl

and formaldehyde, the reactions to diazolidinyl urea

were probably caused by formaldehyde in the patch

test (5). Such studies, of which there are many,

indicate that at least in a number of patients co-

reactions between formaldehyde and releasers are

due to formaldehyde sensitivity. On the other hand,

the investigation of Perret and Happle also indi-

cates that sensitivity to diazolidinyl urea can exist

independent of formaldehyde release, as none of the

seven patients allergic to diazolidinyl urea but not

to formaldehyde reacted to any of other releasers.

The existence of allergy to the releaser 2-

bromo-2-nitropropane-1,3-diol per se (independent

of formaldehyde contact allergy) was demonstrated

25 years ago by Storrs and Bell (6). These inves-

tigators described seven patients reacting to patch

tests with 2-bromo-2-nitropropane-1,3-diol 0.25%

and/or 0.5% and/or 1% aqua or pet. They had

all been sensitized by Eucerin ® cream, which was

widely used in their geographic area as an emollient

for dry and damaged skin, containing 0.05% 2-

bromo-2-nitropropane-1,3-diol. The Eucerin ® cream

was also patch test positive in all patients. None of

them were allergic to formaldehyde and they did not

react to either quaternium-15 or imidazolidinyl urea.

All patients performing a ROAT for a maximum

of 2 weeks on normal skin with Eucerin ® cream

developed dermatitis. This indicated that 2-bromo-

2-nitropropane-1,3-diol allergy per se exists and that

0.05% 2-bromo-2-nitropropane-1,3-diol in a cream

is sufﬁcient to give a positive patch test and a

positive ROAT in 2-bromo-2-nitropropane-1,3-diol

allergic patients who are not sensitive to formalde-

hyde. In addition, the investigators had seen another

four patients reacting to 2-bromo-2-nitropropane-

1,3-diol but NOT to Eucerin ® cream. These patients

DID co-react to formaldehyde and to quaternium-15

and/or imidazolidinyl urea. They concluded that the

amount of formaldehyde in the cream containing

0.05% 2-bromo-2-nitropropane-1,3-diol was insufﬁ-

cient to cause a positive patch test (6).

How often are patients reacting to a releaser also

allergic to formaldehyde?

In many studies, patients have been patch tested with

formaldehyde and one or more releasers, especially

with quaternium-15, which is included in the base-

line series in both Europe and the USA. This gives

an opportunity to study the possible relationship

between formaldehyde and the donors, at least in

this diagnostic setting. In Table 1, data are presented

on the frequency of co-reactions to formaldehyde in

patients allergic to formaldehyde-releasers.

There is a consistent picture for 2-bromo-2-

nitropropane-1,3-diol. Less than 25% (and usually

far less) of 2-bromo-2-nitropropane-1,3-diol-sen-

sitive patients co-react to formaldehyde. This

suggests that most patch test reactions to 2-bromo-

2-nitropropane-1,3-diol indicate sensitivity to this

Table 1. Percentages positive to formaldehyde in patients reacting to formaldehyde-releasers a

2-Bromo-2-nitro-propane-1,3-diol Diazolidinyl urea

DMDM hydantoin

Imidazolidinyl urea

Quaternium-15

Reference

Country and

period of study

Number of

patients

formaldehyde

Number of

patients

formaldehyde

Number of

patients

formaldehyde

Number of

patients

formaldehyde

Number of

patients

formaldehyde

Finland 2001 – 2007

83%

(7)

UK 2004 – 2005

17%

55%

53%

129

59%

(8)

USA 1992 – 2004

788 b

69% b

515 b

83% b

756 b

63% b

(9)

Germany 1994 – 2004

180

37% (?)

(10)

IVDK 1992 – 1995

160

129

12%

15%

47%

(11)

Austria 1992 – 1993

11%

27%

(12)

UK 1982 – 1993

25%

156

24%

299

55%

(13)

USA 1989 – 1991

81%

(14)

USA 1982 – 1989

244

20%

46%

181

65%

(15)

The Netherlands 1984 – 1988

46%

(5)

UK 1985

38%

(3)

UK 1983 – 1984

53%

(3)

USA 1983 – 1987

55%

(15)

USA < 1980

83%

(16)

a Only studies with at least 10 positive reactions to the releasers are included. Data on < 10 patients are presented in Refs (2, 6, 7, 17 – 21).

b Mean numbers/percentages in two patient groups reacting to the releaser in water and to the releaser in pet. vehicle.

Contact Dermatitis 2010: 62: 18–31

FORMALDEHYDE-RELEASERS IN COSMETICS: PART 2 21

Table 2. Percentages positive to formaldehyde-releasers in patients allergic to formaldehyde a

Percentage of patients allergic to formaldehyde with positive reactions to:

Country and

period of study

Number of

patients allergic

to formaldehyde

2-Bromo-2-

nitropropane-

1,3-diol (%)

Diazolidinyl

urea (%)

DMDM

hydantoin

(%)

Imidazolidinyl

urea (%)

Quaternium-15

(%)

Reference

Finland 2001 – 2007 73 – 79

(7)

UK 2004 – 2005

142

(8)

USA 1992 – 2004

2225

24.50

(9)

UK 1992 – 1993

629

(13)

Austria 1992 – 1993

105

(12)

USA 1982 – 1989

454

(4)

USA < 1980

(16)

a Only studies with at least 20 positive reactions to formaldehyde are included. Data on < 20 patients are presented in Refs (2) and (6).

Table 3. Testing with DMDM hydantoin and its parent compound in patients allergic to formaldehyde (22)

Group I. 14 patients tested with DMDM hydantoin

Concomitant allergy to DMDM hydantoin

Patients allergic to formaldehyde

3.0% (only)

1% (also)

0.3% (also)

Negative

1% only

1% and 0.3%

= 4

1% and 0.3% and 0.1%

Group II. 21 patients tested with MDM hydantoin

Concomitant allergy to MDM hydantoin

Patients allergic to formaldehyde

3.0% (only)

1% (also)

0.3% (also)

Negative

1% only

= 12

1% and 0.3%

1% and 0.3% and 0.1%

preservative per se and are not related to formalde-

hyde allergy. The picture for quaternium-15 is also

fairly consistent. In all but two studies, half the

number of patients or more reacted to formalde-

hyde, indicating that formaldehyde is in > 50% of

the cases responsible for the quaternium-15 patch

tests. Less consistent are the ﬁgures for diazo-

lidinyl urea (12 – 81% formaldehyde co-reactions),

DMDM hydantoin (37 – 83%), and imidazolidinyl

urea (11 – 63% co-reactions to formaldehyde).

1 molecule

formaldehyde), the other 14 also with DMDM

hydantoin (two molecules formaldehyde). Test con-

centrations were 0.1 – 0.3 – 1% for formaldehyde,

0.3–1–3% for DMDM hydantoin and MDM hydan-

toin, and 1–3–10% for DM hydantoin, all in water

(w/w). Control tests in 97 patients had excluded irri-

tancy. The results are shown in Table 3. No patient

reacted to the parent compound DM hydantoin. Of

the 14 patients in Group I, 8 (57%) reacted to

DMDM hydantoin. Of the 21 patients in Group II, 7

(33%) reacted to MDM hydantoin. In both groups,

most negative reactions to (D)MDM hydantoin were

observed in patients reacting only to the 1% solution

of formaldehyde. Patients with ‘stronger’ allergies to

formaldehyde (reacting to 0.1% and/or 0.3% also)

tended to show more positive reactions even to the

lower concentrations of (D)MDM hydantoin.

How often are patients reacting to formaldehyde

also allergic to one or more

formaldehyde-releasers?

The reverse situation, i.e. what percentage of

patients allergic to formaldehyde reacts to the

releasers, can also be deducted from many studies.

The available data are shown in Table 2.

For 2-bromo-2-nitropropane-1,3-diol, the data are

consistent: less than 10% of patients allergic to

formaldehyde react to 2-bromo-2-nitropropane-1,3-

diol. On the other side of the spectrum, in all but

two studies at least 30% of the patients reacting to

formaldehyde also reacted to quaternium-15. Less

patients ( < 30%) will react to diazolidinyl urea, and

still fewer ( < 23%) to imidazolidinyl urea. One study

showed that about one out of ﬁve formaldehyde-

sensitive individuals was also patch test positive to

DMDM hydantoin. The relationship between posi-

tive patch tests to formaldehyde allergy and patch

test reactions to DMDM hydantoin was investigated

in detail by De Groot et al (22). Thirty-ﬁve patients

allergic to formaldehyde were patch tested with

formaldehyde and dimethylhydantoin (DM hydan-

toin; the parent compound). Twenty-one of these

formaldehyde-sensitive subjects were also tested

with MDM hydantoin (DM hydantoin

22 DE GROOT ET AL.

Contact Dermatitis 2010: 62: 18–31

Table 4. Percentages positive to other formaldehyde-releasers in patients allergic to a particular formaldehyde-releaser

Percentage of patients allergic to releaser with positive reactions to:

Country and

period of study

Number of

patients allergic

to formaldehyde

2-Bromo-2-

nitropropane-

1,3-diol (%)

Diazolidinyl

urea (%)

Imidazolidinyl

urea (%)

Quaternium-15 (%)

Reference

2-Bromo-2-nitropropane-1,3-diol

UK 2004 – 2005

(100)

(8)

UK 1982 – 1993

(100)

(13)

USA 1982 – 1989

244

(100)

(4)

Diazolidinyl urea

UK 2004 – 2005

(100)

(8)

UK 1982 – 1993

(100)

(13)

USA 1981 – 1991

(100)

(14)

Imidazolidinyl urea

UK 2004 – 2005

(100)

(8)

UK 1982 – 1993

156

(100)

(13)

USA 1982 – 1989

(100)

(4)

Quaternium-15

UK 2004 – 2005

129

(100)

(8)

UK 1982 – 1993

154

(100)

(13)

USA 1982 – 1989

181

(100)

(4)

NT, not tested.

Using the data from Rosen and McFarland (23),

it was estimated that 3% DMDM hydantoin con-

tains 0.3% (3000 p.p.m.) and the 0.3% solution

0.02% (200 p.p.m.) free formaldehyde. Thus, 3 of

14 patients reacted to 200 p.p.m. upon patch testing.

This may be comfounded somewhat by the fact that

the patients were also concurrently tested with the

higher concentrations and with formaldehyde (addi-

tive effect). These data demonstrate that aqueous

preparations of DMDM hydantoin, in concentrations

comparable to those used in cosmetic products, con-

tain enough free formaldehyde to cause dermatitis in

some patients allergic to formaldehyde when patch

tested (22).

a consistent pattern: few reactions to and from other

formaldehyde-releasers, formaldehyde plays a minor

role. There is in two of three studies a very high

co-reactivity between imidazolidinyl urea and dia-

zolidinyl urea, higher than with formaldehyde, indi-

cating (true) cross-allergy or a reaction to a common

(non-formaldehyde) constituent (see under imidazo-

lidinyl urea and under diazolidinyl urea, Ref. (1)).

Of the patients reactive to diazolidinyl urea or imi-

dazolidinyl urea, up to nearly half have reactions to

quaternium-15, indicating a role for formaldehyde

because these chemicals are structurally unrelated.

Conversely, however, of the patients reacting to

quaternium-15, only up to 23% react to imidazo-

lidinyl urea and up to 30% to diazolidinyl urea.

In such cases, the patient may well be allergic to

formaldehyde, but the amount of formaldehyde in

the diazolidinyl urea and imidazolidinyl urea test

substances is not enough to cause a positive patch

test reaction.

The recent experience of the British Contact

Dermatitis Society with formaldehyde and formal-

dehyde-releasers is shown in Table 5 (8). This table

How often are patients reacting to a formaldehyde

donor also allergic to one or more other

formaldehyde-releasers?

In four large studies, the dynamics of patch test reac-

tivity between the various formaldehyde-releasers

was studied. The results are shown in Table 4.

Again, for 2-bromo-2-nitropropane-1,3-diol, there is

Table 5. Relationship between formaldehyde and formaldehyde-releasers: UK experience 2004 – 2005 (8))

Formaldehyde (%) Quaternium-15 (%)

2-Bromo-2-nitro-

propane-1,3,diol (%)

Imidazolidinyl

urea (%)

Diazolidinyl urea (%)

Formaldehyde

(100)

Quaternium-15

(100)

2-Bromo-2-nitropropane-

(100)

1,3-diol

Imidazolidinyl urea 53 47 12 (100) 75

Diazolidinyl urea 55 49 11 60 (100)

This table may be read as follows: horizontal row, 52% of patients allergic to formaldehyde are also allergic to quaternium-15 and 23%

are also allergic to imidazolidinyl urea. Conversely, vertical row, 59% of patients allergic to quaternium-15 are allergic to formaldehyde,

and 53% of the patients allergic to imidazolidinyl urea (55% for diazolidinyl urea) are also allergic to formaldehyde.

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: