oo ORAL SURGERY

Vol. 95 No. 3 March 2003

o ORAL MEDICINE
ORAL PATHOLOGY

ORAL MEDICINE Editor: Martin S. Greenberg

Amalgam-contact hypersensitivity lesions and oral lichen planus

Martin H. Thornhill, MBBS, BDS, PhD,a Michael N. Pemberton, MBChB, BDS,b
Raymond K. Simmons, DDS, DMSc,c and Elizabeth D. Theaker, BDS, MPhil,d San Antonio,
Tex, and Manchester, England
UNIVERSITY OF TEXAS HEALTH SCIENCE CENTER SAN ANTONIO AND UNIVERSITY DENTAL HOSPITAL OF
MANCHESTER

Objective. The purpose of this study was to investigate the relationship between amalgam restorations and oral lichen
planus.
Study design. Eighty-one patients with oral lichenoid lesions were characterized clinically and skin patch tested for
amalgam or mercury hypersensitivity. Thirty-three of these patients had amalgam fillings in contact with oral lesions
replaced and were followed to determine the outcome.
Results. Clinically, 2 patient groups were identified: (1) 30 patients with probable amalgam-contact hypersensitivity
lesions (ACHLs) and (2) 51 patients with oral lichen planus (OLP) but no clear relationship with amalgam. Seventy
percent of ACHL cases were patch test positive for amalgam or mercury compared with only 3.9% of OLP cases (P ⬍
.0001). Amalgam replacement resulted in lesion improvement in 93% of ACHL cases. No such improvement was
observed in the OLP cases treated (P ⬍ .001).
Conclusion. OLP is a heterogeneous condition within which an ACHL subgroup can be identified. ACHLs, but not
other OLP lesions, respond favorably to amalgam replacement. A strong clinical association between lesions and
amalgam restorations plus a positive patch test result was a good predictor of lesion improvement on amalgam
replacement.
(Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2003;95:291-9)

Amalgam is the most widely used dental restorative
material in the world. However, because of the contin-
uous low-level release of mercury from amalgam fill-
ings, its safety and wide scale use have been ques-
tioned.1-5 The main concerns relate to (1) the potential
toxic effects of mercury and (2) the possibility that
mercury may induce adverse immunological effects.
With regard to the latter, there have been a number of
a
Chair and Professor, Department of Dental Diagnostic Science,
University of Texas Health Science Center San Antonio.
b
Consultant in Oral Medicine, Unit of Oral Medicine, University
Dental Hospital of Manchester.
c outcome: cell-mediated autoimmune damage to basal
Assistant Professor, Department of Dental Diagnostic Science, Uni-
versity of Texas Health Science Center San Antonio.
d
Lecturer, Unit of Oral Medicine, University Dental Hospital of
Manchester.
Received for publication Aug 16, 2002; returned for revision Oct 7,
2002; accepted for publication Dec 4, 2002.
© 2003, Mosby, Inc.
1079-2104/2003/$30.00 ⫹ 0
doi:10.1067/moe.2003.115
reports suggesting that amalgam fillings may induce
oral lichen planus (OLP) or oral lichenoid lesions.6-13
However, the precise relationship is not clear. Skin
patch test studies to investigate contact sensitivity re-
sponses to mercury and amalgam have produced con-
flicting results with between 8% and 78.9% of OLP
patients being positive.12-19 Studies investigating the
effect of removing amalgam fillings in OLP patients
have also produced variable results.12,13,15,18-21
OLP is a common condition with a prevalence of 2%
in the general population.22,23 However, there is a
growing recognition that the lesions of OLP may result
from a number of different causes with a common
oral keratinocytes.24 It is postulated that basal keratin-
ocytes may become the target for cell-mediated auto-
immune damage when they express altered or foreign
antigens on their surface. Classically, OLP lesions are
bilateral and symmetrically distributed, suggesting that
systemic factors or systemically distributed antigens
play a role in altering the antigenicity of basal keratin-

291
292 Thornhill et al ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY
March 2003

Table I. Grading of strength of association between mucosal lesions and amalgam restorations
Grade
1. No association
2. Weak association
3. Strong association
4. Very strong association
Criteria
No lesions in direct contact with amalgam restorations. Lesions have typical bilaterally symmetrical
distribution of oral lichen planus or are restricted to areas of the palate, gingivae, or outer
surface of the lip that do not come into contact with the teeth.
Some amalgam restorations in contact with affected areas of mucosa. However, ⬍25% of affected
mucosa in direct contact with amalgam restorations. Some amalgam restorations may also be in
direct contact with unaffected areas of oral mucosa.
⬎75% of affected mucosa in direct contact with amalgam restorations. No amalgam restorations in
direct contact with unaffected areas of oral mucosa.
Lesions restricted to but affecting all areas of mucosa in direct contact with amalgam restorations.

ocytes. In most cases, however, the precise nature of
the responsible antigen is not known.
In contrast, amalgam fillings are in direct contact
with the oral mucosa and may directly alter the antige-
nicity of basal keratinocytes by the release of mercury
and other metal salts as corrosion products.7,9,24 In
susceptible individuals, therefore, amalgam fillings
may induce amalgam-contact hypersensitivity lesions
(ACHLs) with features similar to OLP.8,9,24 Such le-
sions are likely to occur on mucosal surfaces in intimate
contact with amalgam fillings and could be expected to
improve following removal of the fillings.
Our hypothesis was that in individuals sensitized to
mercury or amalgam alloy, contact hypersensitivity le-
sions with features similar to OLP may develop in areas
of the oral mucosa in direct contact with amalgam
restorations, but that in other patients with classical
features of OLP, hypersensitivity to mercury or amal-
gam plays no role.
The purpose of this study was, therefore, to clarify
the relationship between amalgam fillings, OLP, and
possible ACHL; to identify features that help to distin-
guish cases in which amalgam is likely to play a causal
role; and to evaluate the benefit of amalgam removal in
those cases.
MATERIAL AND METHODS
Patients
Eighty-one dentate patients with at least 1 amalgam
restoration who were referred to the University Dental
Hospital of Manchester with a clinical diagnosis of
OLP or oral lichenoid reaction to amalgam were in-
cluded in this study. Diagnosis was based on the clin-
ical criteria described by Eisen.25 These include
(1) reticular (net/plaque-like white patches), (2) ero-
sive/atrophic (erythematous areas of thinned but unbro-
ken epithelium, including desquamative gingivitis) and
(3) ulcerative (with broken epithelium) lesions. In each
case the clinical diagnosis was confirmed by a biopsy
with a histologic appearance that included a band-like,
mainly lymphocytic, immunoinflammatory infiltrate in
the connective tissue adjacent to the epithelial basement
membrane; liquifaction degeneration of the basement
membrane; and destruction of basal keratinocytes.
A standardized history and oral examination was
recorded for each patient. The clinical details of lesions
were recorded and photographed to aid assessment of
any change. Lesions were described as reticular, or
plaque-like, if the clinical appearance was mainly hy-
perkeratotic. Lesions were described as erosive if they
exhibited erythematous change and ulcerated if there
was frank ulceration. The presence of desquamative
gingivitis was recorded separately, as was a history of
skin lesions consistent with OLP. A dental charting was
performed in which particular note was made of any
amalgam restorations in direct physical contact with the
buccal/labial mucosa (buccal contacts) or lingual mu-
cosa (lingual contacts) with the jaws at rest or during
normal physiological movement. Patients were graded
clinically on a 4-point scale (Table I) on the strength of
association between their mucosal lesions and their
amalgam restorations. Grades were assigned by at least
2 clinicians. In most cases there was agreement about
the assigned grade but, when there was a disparity, a
grade was decided by discussion between the clini-
cians. Note was also made of the presence of dissimilar
metals, eg, amalgam and gold used as restorations in
contacting adjacent or opposing teeth.
All patients were skin patch tested in the Oral Med-
icine Clinic at the University Dental Hospital of
Manchester by a physician experienced in skin patch
testing. The following allergens were used: 1% ammo-
niated mercury in petrolatum (Trolab allergen E0602;
Trolab Biodiagnostics, Worcestershire, United King-
dom), 5% amalgam in petrolatum (Trolab allergen
E2509), and 20% amalgam alloying metals in petrola-
tum (Trolab allergen E2508), along with other mem-
bers of the European Standard Series and Dental Ma-
terials Series of patch test allergens (Trolab).26,27
Patients were considered to be patch test positive to an
allergen if they developed a weak positive (nonvesicu-
lar, erythematous reaction), strong positive (edematous
ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY Thornhill et al 293
Volume 95, Number 3

Table II. Grading of lesion improvement following removal of amalgam restorations

Grade
1. No improvement
2. A little improvement
3. Substantial improvement
4. Complete improvement
Criteria
No improvement in the overall size or severity of the lesions.
⬍50% reduction in the overall size of the lesions or ⬍50% reduction in the size of any erosive
or ulcerated areas.
⬎50% reduction in the overall size of the lesions plus ⬎50% reduction in the size of any
erosive or ulcerated areas.
Complete resolution of the lesion.

or vesicular reaction), or a very strong positive (bullous
or ulcerative) hypersensitive skin response at the site of
contact following 72 hours exposure to the allergen.26,27
Very weak or doubtful responses and responses that
were considered to be irritant in nature were regarded
as negative.27 In addition, all patients were biopsied and
the material was sent for routine histologic examina-
tion. Ethics approval for this study was obtained from
the Manchester Health Commissions Research Ethics
Committee (Central). Biopsies were routinely reported
without knowledge of the results of skin patch testing.
The routine diagnostic pathology reports for the biop-
sies from these patients were reported by the patholo-
gist, using standard diagnostic criteria, as consistent
with OLP, consistent with OLP but with some lichen-
oid features, or consistent with a lichenoid reaction.28
For statistical analysis the last 2 categories were
grouped together.
All patients who had a strong or very strong clinical
association (Table I) between their amalgam restora-
tions and their oral lesions were advised to have amal-
gam restorations replaced. In each case, replacement
was carried out by the patient’s own general dental
practitioner. Specific written guidelines were given to
the patient’s dentist. Dentists were advised to replace
only amalgam restorations in direct contact with af-
fected areas of oral mucosa and to use glass ionomer or
composite filling materials or gold, ceramic, or ceramic
bonded to precious metal crowns, according to clinical
need. When crowns were used, dentists were advised
that the amalgam could be retained as a core only when
it was entirely covered by the crown. Patients and their
dentists were followed to ascertain the precise nature of
all replacement restorations and the effect replacement
had on the lesions. Any improvement in the lesions was
graded on a 4-point scale (Table II) by comparison with
photos of the lesions before amalgam replacement.
relating to the degree of clinical improvement follow-
ing removal of amalgam fillings were dichotomized.
Thus strong and very strong clinical associations be-
tween lesions and amalgam fillings were grouped as
strong association, and weak associations or no asso-
ciation were grouped as weak. Similarly, patients with
complete or substantial improvement in lesions follow-
ing amalgam removal were grouped as improved, and
those with little or no improvement were grouped as no
significant improvement. These data were analyzed
with ␹2 cross tabulation using the Fisher exact test.
In order to assess the efficacy of amalgam replace-
ment, we calculated, when appropriate, the number of
patients who needed to be treated to obtain 1 successful
outcome, ie, the number-needed-to-treat (NNT) value.
A successful outcome was defined as complete or sub-
stantial lesion improvement (Table II). The closer the
NNT value was to 1, the more successful the treatment;
the larger the value, the less successful the treatment.
We also looked at how effective strength of clinical
association alone, skin patch test positivity, or a com-
bination of the 2 was in predicting which patients
would benefit from amalgam replacement treatment.
Because not all patients in each group had their amal-
gams replaced (2 declined treatment, and we did not
have approval for 2 patch test–positive patients with
weak or no clinical association between their lesions
and their amalgam restorations), we assumed for sta-
tistical purposes that patients who did not have amal-
gams replaced would not have benefited from the pro-
cedure (worst-case scenario). The result of this
assumption is an underestimate of the predictive value
of the criteria being tested. The odds ratio was also
calculated in order to compare the predictive value of
clinical criteria alone, skin patch testing alone, or a
combination of the 2 in identifying patients who would
benefit from amalgam replacement therapy.

Statistics
Data were converted for cross-tabulation statistical
analysis by using the ␹2 test. For further analysis, the
data relating to the strength of association between
mucosal lesions and amalgam fillings and the data
RESULTS
Eighty-one patients with a mean age of 54.6 years
(Table III) were recruited for the study: 21 men
(25.9%) and 60 women (74.1%). Twenty-three (28.4%)
of these patients were patch test positive for mercury or
294 Thornhill et al ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY
March 2003

Table III. Age and sex details of the different patient groups
All Female Male
Age Age Age
Patient group n (%)* mean ⫾ SD range n (%) †
mean ⫾ SD range n (%) †
mean ⫾ SD range
All patients 81 (100%) 54.6 ⫾ 11.0 28-75 60 (74.1%) 54.9 ⫾ 10.5 28-74 21 (25.9%) 53.9 ⫾ 12.5 30-75
Response to skin patch testing
Patch test ⫹ 23 (28.4%) 53.0 ⫾ 9.2 33-73 17 (73.9%) 53.1 ⫾ 9.0 33-73 6 (26.1%) 52.7 ⫾ 10.5 38-70
Patch test – 58 (71.6%) 55.2 ⫾ 11.6 28-75 43 (74.1%) 55.5 ⫾ 11.1 28-74 15 (25.9%) 54.3 ⫾ 13.5 30-75
Clinical association with amalgam
Very strong 25 (30.9%) 58.8 ⫾ 10.8 28-73
Strong 5 (6.2%) 52.4 ⫾ 11.6 33-63
Weak 21 (25.9%) 52.8 ⫾ 11.0 30-70
None 30(37.0%) 57.7 ⫾ 10.9 31-75
V strong ⫹ strong 30 (37.0%) 52.8 ⫾ 10.7 28-73 22 (36.7%) 52.5 ⫾ 11.2 28-73 8 (38.1%) 53.7 ⫾ 10.0 38-70
Weak ⫹ none 51 (63.0%) 55.7 ⫾ 11.1 30-75 38 (63.3%) 56.6 ⫾ 10.0 34-74 13 (61.9%) 53.1 ⫾ 14.1 30-75

*Percentages in this column are percentages of the total population (n ⫽ 81).

†
Percentages in these columns are the percentages of female and male patients respectively in each patient group.

amalgam. There was no significant difference in the age
or sex distribution among those who were patch test
positive, patch test negative, or the total population
(Table III). There was also no significant difference in
the age or sex distribution of the groups with different
strengths of clinical association between their amalgam
restorations and mucosal lesions.
Of those patients who had a strong or very strong
clinical association between their amalgam restorations
and mucosal lesions (n ⫽ 30), 21 (70%) were patch test
positive for mercury or amalgam. Thirteen were posi-
tive for mercury alone, 1 for amalgam alone, and 7 for
both. In contrast, significantly fewer (2 out of 51
[3.9%]) of those with weak or no association between
their fillings and their mucosal lesions were patch test
positive (P ⬍ .00001). One of these patients reacted to
mercury, and the other reacted to both mercury and
amalgam. All of those patients who were patch test
positive for amalgam (23) reacted to 5% amalgam in
petrolatum (Trolab allergen E2509). Only 1 was addi-
tionally positive to 20% amalgam alloying metals (Tro-
lab allergen E2508).
Reticular, erosive, and ulcerative lesions were
seen in all groups of patients. In contrast, desquama-
tive gingivitis was not observed in any of the patch
test–positive patients or patients with a strong or
very strong clinical association between their lesions
and their fillings. Desquamative gingivitis was, how-
ever, common in patients who were patch test neg-
ative or had weak or no clinical association between
the oral lesions and their amalgam fillings (P ⬍ .001
in both cases; Table IV). Similarly, none of the patch
test-positive patients or patients with a very strong or
strong clinical association between their lesions and
their amalgam fillings gave a history of having skin
lesions, whereas skin lesions had been present at
some stage in a significant proportion of those who
were patch test negative or had weak or no clinical
association between the oral lesions and their amal-
gam fillings (P ⬍ .05 in both cases; Table IV).
Erosive lesions occurred with greater frequency in
patients who were patch test positive for mercury or
amalgam or who had a strong clinical association
between their lesions and their amalgam fillings.
However, only in the latter case was this difference
statistically significant (P ⬍ .01).
Significantly more patients who were patch test
positive for mercury or amalgam had fillings that
were in direct contact with the buccal mucosa (P ⬍
.01), either the buccal or lingual mucosa (P ⬍ .01),
or both (P ⬍ .05) than patients who were patch test
negative (Table V).
History of atopy
Overall, 32% of those in the study had a positive
history of atopy (eczema, asthma, hay fever, or allergy
to medicines, foods, etc). However, there was no sig-
nificant association between a history of atopy and
patch test positivity for mercury/amalgam or the devel-
opment of lichenoid mucosal lesions associated with
amalgam fillings.
Biopsy results
The routine pathology reports for the biopsies
from these patients were reported as consistent with
OLP (n ⫽ 60), consistent with OLP but with some
lichenoid features (n ⫽ 10), or consistent with a
lichenoid reaction (n ⫽ 11; Table VI). For statistical
analysis, the last 2 categories were grouped together.
There was no significant correlation between the
ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY Thornhill et al 295
Volume 95, Number 3

Table IV. Frequency of different types of clinical lesion in the different patient groups
Lesion type
Desquamative History of
Patient group Reticular Erosive Ulcerative gingivitis skin lesions
All patients (n ⫽ 81) 79 (97.5%) 44 (54.3%) 13 (16.0%) 23 (28.4%) 8 (9.8%)
Response to skin patch testing
Patch ⫹ (n ⫽ 23) 23 (100%) 15 (65.2%) 4 (17.4%) 0 (0%)* 0 (0%)†
Patch ⫺ (n ⫽ 58) 56 (96.6%) 29 (50%) 9 (15.5%) 23 (39.7%) 8 (13.8%)
Clinical assoc with amalgam
V strong ⫹ strong 30 (100%) 22 (73.3%)‡ 4 (13.3%) 0 (0%)§ 0 (0%)储
(n ⫽ 30)
Weak ⫹ none 49 (96.1%) 22 (43.1%) 9 (17.64%) 23 (45.1%) 8 (15.7%)
(n ⫽ 51)

More than 1 type of lesion may have been present in any 1 patient. Hence percentages for lesion types within any 1 patient group do not total 100. The
frequency of lesions was significantly different between patch test–positive and patch test–negative patients for *desquamative gingivitis (P ⬍ .001) and
†
history of lichen planus skin lesions (P ⬍ .05). The frequency of lesions was also significantly different between patients when there was a very strong or
strong clinical association between their lesions and amalgam fillings and between patients with weak or no association for §desquamative gingivitis (P ⬍
.001), ‡erosive lesions (P ⬍ .01) and 储history of lichen planus skin lesions (P ⬍ .05).

Table V. Presence of amalgam fillings in direct contact with mucosal surfaces in the different patient groups
Presence of amalgam fillings in direct contact with buccal or lingual mucosal surfaces
Patient group Buccal Lingual Both Either/or
All patients (n ⫽ 81) 63 (77.8%) 46 (56.8%) 42 (51.9%) 67 (82.7%)
Response to skin patch testing
Patch ⫹ (n ⫽ 23) 23 (100%)* 16 (69.6%) 16 (69.6%)† 23 (100%)‡
Patch ⫺ (n ⫽ 58) 40 (69.0%) 30 (51.7%) 26 (44.8%) 44 (75.9%)
Clinical assoc with amalgam
V strong ⫹ strong 28 (93.3%)§ 20 (66.7%) 18 (60.0%) 30 (100%)储
(n ⫽ 30)
Weak ⫹ none 35 (68.6%) 26 (51.0%) 24 (47.1%) 37 (72.6%)
(n ⫽ 51)

When patch test–positive and patch test–negative patients were compared, there was a significant difference in the frequency of amalgam fillings that were in
direct contact with *the buccal mucosa (P ⬍ .01), ‡either the buccal or lingual mucosa (P ⬍ .01) or †both (P ⬍ .05). There was also a significant
difference (P ⬍ .01) in the proportion of amalgam fillings in direct contact with §the buccal mucosa or 储the buccal and lingual mucosa between patients where
there was a very strong or strong clinical association between their fillings and their oral lesions and those where there was not.

pathologic diagnoses and the strength of association
between the patients’ lesions and their fillings or the
patients’ response to patch testing. Similarly, lesions
that improved following removal of associated amal-
gam fillings in patch test–positive individuals were
no more frequently diagnosed histologically as li-
chenoid reactions or as having lichenoid features.
Lesion improvement
Of the 30 patients who had a strong or very strong
association between their amalgam restorations and
their lesions, 28 had closely associated fillings re-
placed. In all cases, 1 or more amalgam restora-
tions— usually occlusal—that were not directly as-
sociated with the lesions were left in place. Two
patients who were asymptomatic chose not to have
their fillings replaced. Of the 28 who had their amal-
gams replaced, 20 (71.4%) had complete resolution
of their lesions during a mean follow-up time of
6.4 ⫾ 2.8 months (range: 3-12 months), 6 (21.4%)
had substantial improvement over a mean follow-up
of 16.3 ⫾ 7.5 months (range: 8-27 months), 1 (3.6%)
had a little improvement (after 15 months of follow-
up), and 1 had no improvement (after 8 months of
follow-up). Of the 30 patients with strong or very
strong clinical association who might have benefited
from amalgam replacement, 26 (86.67%) had it done
and benefited. This results in an NNT to obtain
complete or substantial improvement of 1.15.
Twenty-three patients were patch test positive to
mercury or amalgam, and 21 of these had lesions
with a strong or very strong clinical association with
their amalgam fillings. Twenty of these patients had
their restorations replaced, and 19 (95%) exhibited
296 Thornhill et al
Table VI. Biopsy diagnosis
(a) Lesion association with fillings
No/weak association (51)
Moderate/strong association (30)
(b) Patch test response to mercury or amalgam
– (58)
⫹ (23)
(c) Patch test ⫹ patients whose lesions improved on removal of
associated amalgam fillings (19)
Biopsy diagnosis compared with (a) degree of clinical association between lesions and amalgam fillings, (b) response to skin patch testing to mercury or
amalgam, and (c) patch test–positive patients whose lesions improved on removal of any associated amalgam fillings. Biopsy diagnosis was made blind to the
results of patch testing and was as reported for diagnostic purposes.
complete or substantial improvement in their lesions.
Of the 23 patch test–positive patients who might
have benefited from amalgam replacement, 19
(82.61%) had it done and benefited. This results in an
NNT value of 1.21.
Of the 21 patients who were both patch test posi-
tive and had a strong or very strong clinical associ-
ation who might have benefited from amalgam re-
placement, 19 (90.4%) had the procedure done and
benefited, resulting in an NNT value of 1.10. The
combination of a positive skin patch test response
and a strong or very strong clinical association was a
better predictor of lesion improvement than either a
positive skin patch test response or strong or very
strong clinical association alone, with odds ratios of
2 and 1.46 respectively.
In the 28 patients who had amalgam fillings re-
placed on our recommendation, a total of 128 teeth
were filled with 43 composite and 51 glass ionomer
fillings (some teeth had more than 1 filling). Two
teeth were extracted, 28 were ceramic bonded to
precious metal crowns, and 14 gold crowns were
fitted. No significant difference in outcome was
noted with regard to the type of restoration used.
Amalgam was retained or used to form the core in 30
out of the 42 crowns, and this had no significant
adverse effect on the outcome.
Contrary to our advice, 5 patients who were patch
test negative and had only weak or no association
between their amalgam fillings and mucosal lesions
arranged for their fillings to be replaced. In each
case, these patients had all of their amalgam resto-
rations replaced whether they were in contact with
the oral mucosa or not. Of these patients, 3 (60%)
had no improvement in their lesions after 21 ⫾ 5.2
months follow-up (range: 18-27 months), and 2
(40%) had only a little improvement in their lesions
after 18 and 24 months follow-up (NNT ⫽ ⬁). Sig-
ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY
March 2003
Lichenoid reaction/
lichenoid features Lichen planus
13 (25.5%) 38 (74.5%)
8 (26.6%) 22 (73.4%)
12 (20.7%) 46 (79.3%)
9 (39.1%) 14 (60.9%)
7 (36.9%) 12 (63.1%)
nificantly more patients with a strong or very strong
association between their fillings and mucosal le-
sions had complete or substantial resolution of their
lesions when their fillings were replaced (26/28,
93%) than patients who had a weak or no association
with their fillings (0/5, 0%; P ⬍ .0001).
Dissimilar contacting metal restorations were ob-
served in only 12 of the 81 (15%) patients in this study.
However, there was no significant association between
the presence of contacts and the strength of association
with any mucosal lesions, patch test positivity, or out-
come from amalgam replacement.
DISCUSSION
We have demonstrated that amalgam restorations
may induce lichenoid ACHLs of the oral mucosa in
susceptible individuals. Furthermore, removal of these
fillings results in clinical resolution of the lesions. Al-
though these lesions have a number of the features of
OLP, certain clinical features as well as the results of
skin patch testing against mercury and amalgam can
help to distinguish those patients who are likely to
benefit from amalgam removal from those who will
not.
Clinically, it was possible to divide patients into 2
groups. In the first group were those patients with a
very strong or strong physical association between their
fillings and their mucosal lesions. In the second group
were those patients with only a weak physical associ-
ation or no association between their amalgam fillings
and the more typical bilateral and symmetrical lesions
of OLP.
It was also possible to divide patients into 2 groups
on the basis of their skin patch test response to mercury
or amalgam. Seventy percent of those with a very
strong or strong association between their fillings and
their mucosal lesions produced a positive skin patch
test response to ammoniated mercury or amalgam. This
ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY
Volume 95, Number 3
contrasted with only 3.9% among those with weak or
no association. This figure is not significantly different
from the 3.2% of individuals found to be patch test
positive to mercury in the general population.12,29 All
but 1 of the patients who produced a positive patch test
response did so to ammoniated mercury. One patient
had a positive reaction to amalgam alone, while 8 had
positive reactions for both. In testing for sensitivity to
amalgam, 5% amalgam in petrolatum proved a more
sensitive allergen than the 20% amalgam alloying met-
als. The widely different responses to patch testing in
other studies probably reflects different proportions of
individuals with ACHL or OLP in the studies, since
most studies have not differentiated between these 2
groups.
Although desquamative gingivitis and a history of
skin lesions are found in a proportion of patients with
OLP, neither of these were present in patients with a
very strong or strong association between their amal-
gam fillings and their mucosal lesions. They were not a
feature either in those who were patch test positive for
mercury or amalgam. In contrast, erosive lesions oc-
curred more frequently in patch test-positive patients
and those with a strong clinical association with their
fillings.
On this basis, it was possible to divide oral lesions
with a lichen planus or lichenoid appearance into 2
groups: ACHL and OLP. ACHL and OLP can be
distinguished on clinical grounds by the criteria in
Table I. Grades 1 and 2 favor OLP, and 3 and 4 favor
ACHL. The presence of desquamative gingivitis or skin
lesions further favors OLP. Skin patch testing is also
helpful with a positive reaction for ammoniated mer-
cury or amalgam significantly improving the specific-
ity, if not the sensitivity, of discrimination for ACHL.
In contrast, histopathology of lesional biopsies, while
excellent at distinguishing ACHL and OLP from other
lesions, was not particularly good at distinguishing
between these 2 lesions when performed blind to patch
test information.
Prolonged intimate contact of the oral mucosa with
amalgam fillings in susceptible individuals appears
to be responsible for ACHL. It was notable that the
mean age of patients at presentation was 54.6 years,
and in most cases the mucosally contacting amalgam
fillings had been present for many years before pa-
tients or clinicians had identified the presence of
ACHL. It seems likely that the induction of ACHL is
a slow process, possibly resulting from the release of
mercury salts and other corrosion products from the
surface of the restoration. After crossing the oral
epithelium, these may haptenize with oral keratino-
cyte surface proteins. In susceptible individuals this
may eventually result in lymphocyte activation and
Thornhill et al 297
the induction of a cell-mediated autoimmune re-
sponse directed at basal keratinocytes.24,30-32 Be-
cause all individuals with mucosally contacting
amalgams are exposed in the same way, it is likely
that the difference between responders and non-
responders is genetically determined in some way,
possibly by HLA type. Although individuals with
one type of hypersensitivity often have others, we
found no significant difference in the proportion of
ACHL or OLP patients with an atopic history. How-
ever, atopic responses are generally antibody medi-
ated, and it is likely that ACHL and OLP are cell
mediated.
In the past lichenoid lesions caused by contact
with restorations have been attributed to galvanic
reactions between dissimilar metals in close con-
tact,33-36 eg, the electrical circuit that may exist be-
tween a contacting gold crown and an amalgam
filling in an adjacent tooth. It was postulated that this
circuit induced changes in the adjacent mucosa. To
investigate this hypothesis, we recorded the presence
of dissimilar contacting metal restorations but found
no significant association between them and the pres-
ence of lesions, patch test positivity, or clinical im-
provement on amalgam replacement. Our findings,
therefore, did not support the concept of galvanic
lesions. Rather, these lesions appear to be the result
of cell-mediated contact hypersensitivity responses
to mercury or amalgam in susceptible individuals
who have been sensitized through long exposure.
As for OLP, despite a large amount of research
into the pathology and nature of the condition,24,31,32
the cause is still not clear. The bilateral, symmetrical
distribution of the lesions suggests a systemic cause.
This is particularly so when skin lesions are also
present. However, it is also possible that the sym-
metrical distribution could result from a contact hy-
persensitivity response to antigens that are more
evenly distributed in the mouth. These could be
food- or plaque-related microbial antigens in suscep-
tible, sensitized individuals. With the exception of
specific examples such as ACHL, we do not know
the cause in the majority of patients with OLP.
Indeed, OLP may represent a group of diseases, each
with a different cause but sharing a similar clinical
outcome. ACHL may represent the easiest subgroup
of OLP to distinguish because of its association with
amalgam.
In general, because we can’t treat the underlying
cause of OLP, the approach to treatment is to sup-
press it with immunosuppressant drugs. Cessation of
treatment usually results in reoccurrence. In the case
of ACHL, however, we can treat it by covering or
replacing contacting amalgam fillings. In patients
298 Thornhill et al
who are patch test positive for mercury or amalgam,
or in whom a strong clinical association between
their lesions and restorations exists, such treatment is
highly effective. However, the combination of a pos-
itive patch test response and strong clinical associa-
tion was an even better predictor of lesion improve-
ment, with an NNT of only 1.10. In the majority of
such patients, lesion improvement was gradual but
complete with a mean time to resolution of 6.4
months. Once resolved, new lesions arose only if
new amalgam restorations were placed in contact
with the mucosa. However, once the patient was
sensitized, the placement of a new amalgam could
result in vigorous lesion development. The variable
response to amalgam replacement noted in other
studies probably reflects differing proportions of
ACHL and OLP patients in the study groups.
It was notable that only amalgam fillings in direct
contact with the mucosa needed to be removed to
achieve lesion resolution. The different replacement
filling or crown materials used were equally effec-
tive. However, it became clear that inert materials
are preferable. One patient who had amalgam fillings
replaced with composite showed lesion improvement
initially but then relapsed. Prior to amalgam re-
moval, the patient was patch test positive for mer-
cury but negative for Bis-GMA, a constituent of
composite filling materials. When retested following
relapse, the patient was positive for both. Replace-
ment of the composite restorations with glass iono-
mer resulted in complete resolution of the lesions. It
is possible that the increased permeability of mucosa
affected by erosive ACHLs could facilitate sensiti-
zation to other materials. In view of this, inert ma-
terials such as glass ionomer fillings, porcelain res-
torations, porcelain crowns, or porcelain covered
precious metal crowns are preferable for patients
with ACHL. We did not identify any benefit in
removing or replacing amalgam fillings in patients
with OLP of unknown cause.
Further studies in different population groups are
needed to confirm our findings and further evaluate the
benefit of skin patch testing and amalgam replacement
in OLP and ACHLs.
We would like to thank all those dentists who assisted this
study by carrying out the replacement of amalgam fillings on
their patients and collaborated in providing data and informa-
tion. We would also like to thank colleagues who referred
patients for inclusion in the study and the Unit of Oral
Pathology at the University Dental Hospital of Manchester
for providing diagnostic pathology information on these pa-
tients.
ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY
March 2003
REFERENCES
1. Hanson M, Pleva J. The dental amalgam issue. A review. Expe-
rientia 1991;47:9-22.
2. ADA Council on Scientific Affairs. Dental amalgam: update on
safety concerns. J Am Dent Assoc 1998;129:494-503.
3. Corbin SB, Kohn WG. The benefits and risks of dental amalgam:
current findings reviewed. J Am Dent Assoc 1994;125:381-8.
4. Eley BM. The future of dental amalgam: a review of the litera-
ture. Part 6: Possible harmful effects of mercury from dental
amalgam. Br Dent J 1997;182:455-9.
5. Levy M. Dental amalgam: toxicological evaluation and health
risk assessment. J Can Dent Assoc 1995;61:667-8, 71-4.
6. Mobacken H, Hersle K, Sloberg K, Thilander H. Oral lichen
planus: hypersensitivity to dental restoration material. Contact
Dermatitis 1984;10:11-5.
7. Lundstrom IM. Allergy and corrosion of dental materials in
patients with oral lichen planus. Int J Oral Surg 1984;13:16-
24.
8. Camisa C, Taylor JS, Bernat JR, Jr, Helm TN. Contact hyper-
sensitivity to mercury in amalgam restorations may mimic oral
lichen planus. Cutis 1999;63:189-92.
9. Smart ER, Macleod RI, Lawrence CM. Resolution of lichen
planus following removal of amalgam restorations in patients
with proven allergy to mercury salts: a pilot study. Br Dent J
1995;178:108-12.
10. Ostman PO, Anneroth G, Skoglund A. Oral lichen planus lesions
in contact with amalgam fillings: a clinical, histologic, and im-
munohistochemical study. Scand J Dent Res 1994;102:172-9.
11. Bolewska J, Holmstrup P, Moller-Madsen B, Kenrad B, Dan-
scher G. Amalgam associated mercury accumulations in normal
oral mucosa, oral mucosal lesions of lichen planus and contact
lesions associated with amalgam. J Oral Pathol Med 1990;19:
39-42.
12. Finne K, Goransson K, Winckler L. Oral lichen planus and
contact allergy to mercury. Int J Oral Surg 1982;11:236-9.
13. Koch P, Bahmer FA. Oral lesions and symptoms related to
metals used in dental restorations: a clinical, allergological, and
histologic study. J Am Acad Dermatol 1999;41:422-30.
14. Hietanen J, Pihlman K, Forstrom L, Linder E, Reunala T. No
evidence of hypersensitivity to dental restorative metals in oral
lichen planus. Scand J Dent Res 1987;95:320-7.
15. Laine J, Kalimo K, Forssell H, Happonen RP. Resolution of oral
lichenoid lesions after replacement of amalgam restorations in
patients allergic to mercury compounds. Br J Dermatol 1992;
126:10-5.
16. Skoglund A. Value of epicutaneous patch testing in patients with
oral, mucosal lesions of lichenoid character. Scand J Dent Res
1994;102:216-22.
17. Henriksson E, Mattsson U, Hakansson J. Healing of lichenoid
reactions following removal of amalgam. A clinical follow-up.
J Clin Periodontol 1995;22:287-94.
18. Ibbotson SH, Speight EL, Macleod RI, Smart ER, Lawrence CM.
The relevance and effect of amalgam replacement in subjects
with oral lichenoid reactions. Br J Dermatol 1996;134:420-3.
19. Laine J, Kalimo K, Happonen RP. Contact allergy to dental
restorative materials in patients with oral lichenoid lesions. Con-
tact Dermatitis 1997;36:141-6.
20. Pang BK, Freeman S. Oral lichenoid lesions caused by allergy to
mercury in amalgam fillings. Contact Dermatitis 1995;33:423-7.
21. Bolewska J, Hansen HJ, Holmstrup P, Pindborg JJ, Stangerup M.
Oral mucosal lesions related to silver amalgam restorations. Oral
Surg Oral Med Oral Pathol 1990; 70: 55-8.
22. Axell T. A prevalence study of oral mucosal lesions in an adult
Swedish population. Thesis. Odontol Revs 1976; 27 Suppl 36:
23. Bouquot JE, Gorlin RJ. Leukoplakia, lichen planus, and other
oral keratoses in 23,616 Americans over 35 years of age. Oral
Surg Oral Med Oral Pathol 1986;61:373-81.
24. Thornhill MH. Immune mechanisms in oral lichen planus. Acta
Odontol Scand 2001;59:174-7.
25. Eisen D. The clinical features, malignant potential, and systemic
associations of oral lichen planus: a study of 723 patients. J Am
Acad Dermatol 2002;46:207-14.
ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY Thornhill et al 299
Volume 95, Number 3

26. Brasch J, Geier J, Gefeller O. Dynamic patterns of allergic patch A et al. The pathogenesis of oral lichen planus. Crit Rev Oral
test reactions to 10 European standard allergens. An analysis of Biol Med 2002;13:350-65.
data recorded by the Information Network of Departments of 33. Ullmann K. Leukoplakia caused by electrogalvanic current gen-
Dermatology (IVDK). Contact Dermatitis 1996;35:17-22. erated in the oral cavity. Wien Klin Wochenschr 1932;45:840-4.
27. Gronau U, Treudle K, von Eitzen L-C, Gronau U, Treudle K, von 34. Hollander L. Galvanic burns of the oral mucosa. JAMA 1932;
Eitzen L-CGronau U, Treudle K, von Eitzen L-Cs. The Trolab 99:383-4.
guide to patch testing. Reinbek, West Germany: Hermal Kurt 35. Lain E, Caughron. Electrogalvanic phenomena of the oral cavity
Herrmann; 1987. caused by dissimilar metals. J Am Dent Assoc 1936;23:1641-52.
28. Schiodt M. Oral discoid lupus erythematosus. III. A histopatho- 36. Banoczy J, Roed-Petersen B, Pindborg JJ, Inovay J. Clinical and
logic study of sixty-six patients. Oral Surg Oral Med Oral Pathol histologic studies on electrogalvanically induced oral white le-
1984;57:281-93. sions. Oral Surg Oral Med Oral Pathol 1979;48:319-23.
29. Handley J, Todd D, Burrows D. Mercury allergy in a contact
dermatitis clinic in Northern Ireland. Contact Dermatitis 1993;
29:258-61. Reprint requests:
30. Eversole LR. Immunopathogenesis of oral lichen planus and Martin H Thornhill
recurrent aphthous stomatitis. Semin Cutan Med Surg 1997;16:
Department of Dental Diagnostic Science
284-94.
31. Porter SR, Kirby A, Olsen I, Barrett W. Immunologic aspects of University of Texas Health Science Center San Antonio
dermal and oral lichen planus: a review. Oral Surg Oral Med Oral 7703, Floyd Curl Drive
Pathol Oral Radiol Endod 1997;83:358-66. San Antonio, TX 78229-3900
32. Sugerman PB, Savage NW, Walsh LJ, Zhao ZZ, Zhou XJ, Khan Thornhill@uthscsa.edu

AVAILABILITY OF JOURNAL BACK ISSUES

As a service to our subscribers, copies of back issues of Oral Surgery, Oral Medicine, Oral
Pathology, Oral Radiology, and Endodontics for the preceding 5 years are maintained and are
available for purchase from the publisher, Mosby, until inventory is depleted. The following
quantity discounts are available: 25% off on quantities of 12 to 23; one third off on quantities
of 24 or more. Please write to Mosby, Subscription Customer Service, 6277 Sea Harbor Dr,
Orlando, FL 32887, or call 800-654-2452 or 407-345-4000 for information on availability of
particular issues and prices. If unavailable from the publisher, photocopies of complete issues
are available from Bell & Howell Information and Learning, 300 N Zeeb Rd, Ann Arbor, MI
48106-1346; (734) 761-4700 or (800) 521-0600.