Sensitivity to Implant Materials in Patients Undergoing Total

Hip Replacement
Donatella Granchi,
1
Elisabetta Cenni,
1
Giovanni Trisolino,
2
Armando Giunti,
2
Nicola Baldini
2
1
Laboratory for Pathophysiology of Orthopaedic Implants, Istituti Ortopedici Rizzoli, via di Barbiano 1/10,
40136 Bologna, Italy
2
7th Division of Orthopaedic Surgery, Istituti Ortopedici Rizzoli, via di Barbiano 1/10, 40136 Bologna, Italy
Received 11 April 2005; revised 29 June 2005; accepted 5 July 2005
Published online 1 November 2005 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jbm.b.30445
Abstract: Sensitivity to implant materials is an unpredictable event, which may contribute
to the process leading to the failure of the total hip replacement (THR). The aim of the current
study was to investigate the informative power of skin testing in detecting the sensitization to
the implant components in patients undergoing THR. A consecutive series of 223 patients was
enrolled in the study, including 66 candidates to THR, 53 with stable implant, and 104 with
THR loosening. The patch testing was performed by using the most relevant components of
cobalt-based alloys (CoCrMo), Ti-based alloys (TiAlV), and bone cements. The frequency of
positive patch testing in preimplant patients did not differ from that observed after THR.
Patients with CoCrMo-failed implant showed a signicant low frequency of nickel-positive
skin reaction, while patients with TiAlV-THR had a high incidence of vanadium-positive patch
testing. The panel of haptens showed a good performance in the identication of patients
known to have a contact dermatitis. With regard to the THR outcome, patch testing was not
able to discriminate between stable and failed implant. Sensitivity to at least one hapten,
namely bone cement, as well as the positive medical history of hypersensitivity, inuenced
negatively the THR survival. Our results show the reliability of patch testing for investigating
the sensitivity to implant components. The causeeffect relationship between sensitization and
negative outcome cannot be established, but the shorter lifespan of THR in patients who have
a positive patch testing supports the signicant role of this event in contributing to the implant
failure. 2005 Wiley Periodicals, Inc. J Biomed Mater Res Part B: Appl Biomater 77B: 257264, 2006
Keywords: hip prosthesis; sensitization; cobaltchromium (alloys); titanium (alloys); bone
cement
INTRODUCTION
Materials used for total hip replacement (THR) generally
show a good biocompatibility. The tissue damage following
the surgical implantation does not leave side effects, even if
the host-reactivity to the implant components may inuence
the duration of the event. In this context, the chronic inam-
mation following the generation of wear particles has been
recognized as the main biological mechanism leading to
implant failure. Moreover, if the degradation products are
able to interact with the immune system, adverse immuno-
toxic effect may be induced.
1
The evaluation of immunotox-
icity of orthopedic implant is a prerequisite for its use in
medicine. The Immunotoxicity Testing Guidance issued by
Food and Drug Administration and Center for Devices and
Radiological Health provides an overview of the immuno-
toxic effects that might be associated with bone-implant
materials. Specic immune response, namely hypersensitiv-
ity, immunosuppression, and autoimmunity, but also chronic
inammation and immunostimulation have been considered
as potential adverse effects.
2,3
Metals and polymers represent
the main components of orthopedic implants: in contact with
biological uids they undergo corrosion or degradation, and
ions and molecules that are released may be responsible for
adverse immunotoxic effects.
4
Although little is known about
the bioavailability of circulating degradation products, there
is evidence that novel metalprotein complexes may induce a
delayed-type hypersensitivity reaction (DTH).
5
The metals
known as the most common sensitizers, namely nickel, co-
balt, and chromium, are widely used in orthopedic implants,
but also hypersensitivity to titanium and vanadium have been
described.
6,7
Polymeric biomaterials, namely acrylic bone
cements, are not easily chemically degraded and immuno-
genic reactions to polymethylmethacrylate or other constitu-
ents of the implant have been occasionally reported.
8,9
Sev-
Correspondence to: D. Granchi (e-mail: donatella.granchi.@ior.it)
Contract grant sponsor: Italian Ministry of Health
2005 Wiley Periodicals, Inc.
257
eral cohort studies investigated a possible correlation between
metal-hypersensitivity and implant failure. Despite the heter-
ogeneous patient population, type of the implant, and testing
methodologies, the overall prevalence of metal DTH among
general population was found to be 1015%, approximately
25% in patients with a well-functioning implant, and 60%
among patients with a failed THR.
5
The high proportion of
metal-DTH in patients with implant failure did not prove a
causeeffect relation between these events, and this main
question still remains unsolved.
The assumption of the current study is that the sensitivity
to implant components cannot be considered as unique cause
of implant failure, but a considerable event in the network of
processes promoting the prosthetic loosening, irrespective of
the onset time. Patient undergoing the hip replacement may
be sensitized previously, as well as the sensitization may be
a consequence of the wear of the implant, but in all cases, the
result is a localized immune response that starts or accelerates
the inammatory process leading to the implant failure. In
this context the detection of sensitivity to implant compo-
nents may be considered a remarkable step in the planning of
THR. Several in vitro tests have been addressed in diagnosing
a systemic hypersensitivity to the implant components,
1012
but some faults limit their large-scale application, including
the high costs and the need to be performed in qualied
laboratories. The scarce number of lymphocytes that are
obtained from biological samples does not allow the exhaus-
tive assay of all the immunogenic substances contained in
metal alloys and bone cements; in addition, the toxicity of
some metals cannot be avoided by using the in vitro sys-
tem.
13,14
Alternatively, the in vivo approach, that is the epi-
cutaneous skin testing, may be considered more suitable for
the large-scale application, because it is little expensive and
allows to estimate more haptens simultaneously. This method
is considered the gold standard in diagnosing of contact
allergy, but its validity in detecting a systemic hypersensitiv-
ity is not conrmed.
5
The main goal of the current study was
to elucidate the informative power of the in vivo approach in
detecting a sensitization to the implant components. For this
purpose, we designed a patch testing by using a panel of
haptens chosen among chemicals that may act as sensitizers
in patients undergoing THR, namely metals and cements.
MATERIALS AND METHODS
Study Design
In the period from January 2001 to May 2004, 223 Caucasian
individuals were enrolled in this retrospective case-control
study. The course of the study was authorized from the
Ethical Committee of our institution. A personal history was
obtained from each patient, including details concerning dis-
eases that had led to hip arthroplasty, presence of other bone
or joint implants, other diseases, previous symptoms of con-
tact dermatitis for DTH, previous orthopedic surgery, and
drug intake. Patients with rheumatoid arthritis and those
under corticosteroid or other immunosuppressive treatment
were excluded from the study. Three groups of patients were
evaluated: the rst group (group A) included 66 patients who
were evaluated before THR; group B included 53 patients
with a stable prosthesis (12 months minimum follow-up); and
group C included 104 patients who had clinical and radio-
logical evidence of loosening of THR.
15
Osteoarthritis was
the most frequent indication for THR (65.5%), followed by
congenital hip dysplasia (18.4%), and fractures (10.8%). For-
ty-four patients (19.7%) had an additional bone or joint
implant at another site. Thirty-three patients (14.8%) had a
medical history of metal hypersensitivity, 185 (83%) denied
any kind of previous allergy to metals, and 5 (2.2%) reported
nonspecic symptoms. The type of implant was known in
70.2% of patients: most patients were exposed to both cobalt-
based (CrCoMo) and Ti-based (TiAlV) alloys. Further details
on the clinical series are summarized in Table I.
Patch Testing
Patch testing was carried out by three different observers who
were unaware of clinical information or the composition of
the implant. Hypersensitivity to metals was tested by using
the following haptens: 5% nickel sulphate, 1% cobalt chlo-
ride, 2% chromium trichloride, 0.5% potassium dichromate,
2% ferric chloride, 2% molybdenum chloride, 2% manganese
chloride, 2% vanadium trichloride, 1% aluminium chloride,
and 2% titanium dioxide. The last hapten was chosen as
oxide, rather than as salt, so as to mimic the host reactivity
to titanium-based devices, which form on their surface a
titanium-dioxide lm. Haptens for bone cements were the
following: 5% methyl-methacrylate, 2% butyl-methacrylate,
2% triethylene glycol dimethacrylate, 2% ethylene glycol
dimethacrylate, 2% N,N-dimethyl-paratoluidine, 5% hydroxy-
ethyl-methacrylate, 2% benzoyl-peroxide, and 1% hydroqui-
none monobenzyl ether (F.I.R.M.A SpA, Florence, Italy).
Vaseline, which was the carrier in the patch testing, was
assayed as a negative control. A drop of each hapten was
smeared on the Hayes chamber test, which was applied to
the back of the patient. After 4872 h, skin reactions were
evaluated and graded according to their intensity as 0 (no
reaction), or reaction (week erythema only, doubtful),
1 (erythema with edema covering at least 50% of the patch
test site), 2 (erythema and papules covering at least 50% of
the patch test site), and 3 (vesicles or bullae covering at
least 50% of the patch test site).
16
Calculations and Statistical Analysis
Quantitative results were expressed as mean standard
deviation. The nonparametric analysis of variance (Kruskal-
Wallis) was applied to detect the effects of the various clin-
ical variables on the quantitative results, and the Mann-
Whitney test was applied to detect specic differences be-
tween groups. In each group, the frequency distribution of
positive patch testing was calculated, and the
2
test with
Fishers exact test was used to highlight differences attribut-
able to clinical variables. Probabilistic measures, such as
258 GRANCHI ET AL.
sensitivity, specicity, and likelihood ratios, as well as their
95% condence limits (95% CL), were used to describe the
diagnostic accuracy of our series of haptens in diagnos-
ing DTH patients who had a previously documented metal
hypersensitivity.
17
The same method was employed to
assess the predictive value of patch testing with regard to
the THR status: in this case, specicity showed the effec-
tiveness of the patch testing in excluding a failure; sensi-
tivity showed the effectiveness to detect a failure; and
positive and negative likelihood ratios described the dis-
criminatory properties of positive and negative test results,
respectively. Positive likelihood ratios above 10 and neg-
ative likelihood ratios below 0.1 provide convincing diag-
nostic evidence, whereas those above 5 and below 0.2 give
good diagnostic evidence.
18
The Kaplan-Meier product-limit method was employed to
estimate the 10-year survival rate of the THR, being the
revision considered as the end point of interest; the logrank
Mantel-Cox test was applied to highlight signicant differ-
ences among the survival curves of groups.
17
Differences
were considered signicant if the p value was less or equal to
0.05.
RESULTS
Frequency of Positive Patch Testing
The mean age of group A was signicantly lower than groups
C and B ( p 0.001), but the frequency of skin positive
reaction was not inuenced by age, the indication for THR,
presence of other implants, or presence of documented infec-
tions (data not shown). In all groups, females showed a
signicantly higher incidence of positive patch testing (group
A, 74.2%; group B, 75.5%; and group C, 73.1%). The inci-
dence of positive patch testing to at least one hapten, either
metal or cement, was similar in all groups and no signicant

2
values were found (group A, 43.9%; group B, 50.9%; and
group C, 42.3%). No differences were found with regard to
the intensity of skin reaction: the proportion of 1, 2, and
3 patch testing was similar in all groups, while doubtful
reactions were observed only in three patients (data not
shown).
The incidence of positive patch testing for at least one
metal hapten in preimplant patients (group A) was 39.4%,
and did not differ signicantly from that observed in patients
TABLE I. Prole of the Patients with Loosened Hip Prostheses (group C), Stable Hip Prostheses (group B),
and Pre-Surgery (group A)
Group A (n 66) Group B (n 53) Group C (n 104)
Sex
Males 17(25.8)
a
13 (27.1)
a
27 (25.1)
a
Females 49 (74.2) 35 (72.9) 77 (74.9)
Age (years)
Mean SD 59.6 13 65.0 11 64.7 11
Median 62 68 67
Range 2482 3581 3283
Indication for THR
Idiopathic osteoarthritis 49 (74.2) 33 (62.3) 64 (61.5)
Congenital hip dysplasia 7 (10.6) 15 (28.3) 19 (18.3)
Fractures 0 1 (1.9) 11 (10.6)
Other causes
b
10 (15.2) 4 (7.5) 10 (9.6)
Positive medical history for metal
DTH 12 (18.1) 6 (11.3) 15 (14.4)
Follow-up (months)
Median 76 99
Range 12216 12396
Other implants 10 (18.9) 34 (32.7)
Metal composition of the implant(s)
CoCrMo alloy 2 (3.8) 31 (29.8)
TiAlV alloy 27 (50.9) 25 (25.0)
CoCrMo/TiAlV alloys 24 (45.3) 22 (15.4)
Unknown 0 26 (29.8)
Type of bearing
Ceramic-on-ceramic 47 (88.7) 48 (46.2)
Metal-on-polyethylene 5 (9.1) 39 (37.5)
Metal-on-metal 1 (1.9) 2 (1.9)
Ceramic-on-polyethylene 0 15 (14.4)
Cemented THR 14 (26.4) 50 (48.1)
a
Values in parentheses in these columns indicate percentages.
b
Post-traumatic osteoarthritis (11), osteonecrosis (10), septic osteoarthritis (1), tubercular osteoarthritis (1), Perthes disease (1).
259 SENSITIZATION TO IMPLANT COMPONENTS
of group B (50.9%) and group C (38.5%). No signicant
differences were highlighted when the reactivity against each
hapten was considered, although a common nding was the
higher frequency of positive skin reactions observed in pa-
tients with well-xed implant than that found in patients with
THR failure (Table II).
The type of bearing did not inuence the frequency of
positive skin reaction in group C patients (ceramic-on-ce-
ramic, 57.1%; metal-on-polyethylene, 59.1%; and ceramic-
on-polyethylene, 60.1%). Group B was not considered for the
statistical analysis, because the most part of subjects (88.7%)
had a ceramic-on-ceramic implant.
According to the metal composition of the implant, four
groups were considered: (i) implants with CoCrMo only, (ii)
those with TiAlV only, (iii) those with both alloys (CoCrMo/
TiAlV), and (iv) those with unknown materials. In each group,
the frequency of positive skin reactions to metals was recalcu-
lated (Table III). In the alloys used for the manufacturing of hip
prosthesis, few traces of nickel are present, but since this ele-
ment resulted the most common sensitizer in preimplant patients
(22.7%), the incidence of nickel-positive patch testing was used
as an informative parameter for the immune status of the exam-
ined group. In TiAlV implants, the incidence of metal sensiti-
zation in group C was signicantly higher than in group A; on
the contrary, in patient with failed CrCoMo/TiAlV implants, the
frequency of positive patch testing to at least one metal hapten
decreased, as well as the frequency of nickel-positive reactions;
and the reactivity against nickel decreased also in patients who
had a CoCrMo prosthesis. In group C, who had either TiAlV- or
CoCrMo/TiAlV implant showed a higher incidence of positive
patch testing to vanadium in comparison to individuals of pre-
implant group. In patients with stable THR, the high incidence
of positive skin reaction to vanadium was found only in
CoCrMo/TiAlV group: this nding could be ascribed to the
short follow-up of the patients who had a stable/TiAlV implant.
The incidence of positive reaction to at least one of the
bone cement components (Table IV) was similar in all patient
groups, and no signicant difference was found when posi-
tive skin reaction for each hapten was considered. The fre-
quencies did not change even if the uncemented implants
were considered (data not shown). A further split of the
groups, that is according to the metal composition, was not
performed because of the low frequency of positive patch
testing to bone-cement haptens.
Diagnostic Performance
The ability of patch testing to demonstrate metal DTH in
patients with a positive medical history of contact hypersen-
sitivity was considered (Table V). In the two-by-two cross
table the positive and negative medical history for metal DTH
was matched with the positive and negative patch testing.
TABLE II. Frequency (%) of Positive Skin Reactions to Metal
Haptens in the Patient Groups
Group A
(n 66)
Group B
(n 53)
Group C
(n 104)
At least one hapten 39.4 50.9 38.5
Nickel sulphate 22.7 30.2 15.4
Chromium
a
12.1 9.4 6.7
Cobalt chloride 15.2 15.1 12.5
Ferric chloride 0 1.9 1
Molybdenum chloride 1.5 5.7 1.9
Manganese chloride 10.8 1.9 3.8
Titanium dioxide 1.5 0 1.9
Aluminium chloride 0 1.9 0
Vanadium trichloride 9.1 17.0 15.4
a
Positive patch testing for at least one of the two tested haptens (2% chromium
trichloride or 0.5% potassium dichromate).
TABLE III. Frequency (%) of Positive Skin Reactions to Metal Haptens According to the Metal Composition of the Implant
Group A
(n 66)
Group B
a
(n 53) Group C (n 104)
TiAlV
(n 24)
CoCrMo/TiAlV
(n 27)
CoCrMo
(n 31)
TiAlV
(n 22)
CoCrMo/TiAlV
(n 25)
Unknown
(n 26)
At least one hapten 39.4 41.7 55.6 32.3 68.2
b
28.0
c
30.8
Nickel 22.7 33.3 25.9 12.9 27.3 8.0
d
15.4
Chromium 12.1 8.3 11.1 3.2 9.1 8.0 7.7
Cobalt 15.2 8.3 14.8 12.9 22.7 8.0 7.7
Molybdenum 1.5 4.2 7.4 3.2 4.5 0 0
Titanium 1.5 0 0 0 4.5 0 3.8
Aluminium 0 0 0 0 0 3.7 0
Vanadium 9.1 8.3 25.9
e
3.2 45.5
f,g
23.0
d
0
Follow-up (months;
median) 24 84 132 54 120 54
a
In group B the two patients who had CrCoMo implants were not considered.
b
Group C-TiAlV versus group A, p 0.01.
c
Group C-CoCrMo/TiAlV versus group B-CoCrMo/TiAlV, p 0.05.
d
Group C-CoCrMo/TiAlV versus group A, p 0.05.
e
Group B-CoCrMo/TiAlV versus group A, p 0.04.
f
Group C-TiAlV versus group A, p 0.0005.
g
Group C-TiAlV versus group B-TiAlV, p 0.006.
260 GRANCHI ET AL.
Only two individuals (8.9%), who reported previous symp-
toms of dermatitis because of the contact with jewellery,
showed no positive skin reaction against the tested metals. A
very high sensitivity (0.94; 95% CL: 0.861.02) and negative
likelihood ratio (0.09; 95% CL: 0.020.34) were highlighted,
while specicity (0.67; 95% CL: 0.600.74) and positive
likelihood ratio (2.83; 95% CL: 2.273.52) were not equally
cogent, because a large number of patients with a negative
medical history of DTH and positive patch test was found.
With regards to the status of the THR, that is stable or
failed implant, the probability measurement of diagnostic
accuracy showed that no predictive value may be ascribed to
the patch testing results. In patients who had a skin reactivity
against bone cement a high specicity was found (0.90; 95%
CL: 0.820.98), but the likelihood ratios did not conrm the
discriminatory capability of this test.
Survival Analysis
The median duration of the implant was 120 months in
patients who had no skin reaction, while the THR failure
occurred signicantly earlier in patients with positive patch
testing (78 months; p 0.04), and much more when the
sensitizer was a bone cement hapten (48 months; p 0.01).
No signicant differences were found in patients who were
sensitized against metal haptens (84 months).
Figure 1 shows the Kaplan-Meier survivorship analysis in
patients who had positive or negative patch testing. The
10-year survival rate was 41.3% (95% CL 27.455.2%) in
patients with positive skin reaction to at least one hapten
[Figure (1a)] and 50.5% (95% CL 38.462.5%) in those
who had no evidence of sensitization: the comparison be-
tween the curves showed a statistically signicant difference
( p 0.02). The presence of skin reactions to metal haptens
did not inuence the survival of the implant (positive patch
testing: 44.0%, 95% CL 29.558.5%; negative patch test-
ing: 47.0%, 95% CL 36.259.6%) [Figure (1b)]. On the
contrary, at 10 years none of the patients with positive patch
testing to bone-cement components had a stable THR, while
the survival rate in patients without positive reaction was
50.0% (95% CL 40.259.7; p 0.006) [Figure (1c)].
With regard to the chemical type of sensitizing agent
[Figure (1d)], patients were split in four groups, namely
positive only for metal haptens, only bone cement, both
cement and metal, or negative for all haptens. The statistical
comparison of survival curves showed signicant differences.
The lowest survival rate was observed in patients who were
reactive only to bone cement (0%), followed by the group
with positive reactions against both metal and bone cement
(29.8%, 95% CL 2.257.5%). The survivorship rate of
patients who reacted to metal haptens was similar to that of
the group without skin reaction (48.6%; 95% CL 32.1
65.1% and 49.8%; 95% CL 37.971.8%, respectively).
According to the medical history of metal hypersensitivity,
documented by contact dermatitis or previous patch testing,
three groups were considered [Figure (1e)]. The lowest sur-
vival rate was observed in patients who had a positive patch
testing and a positive medical history (32.0%, 95% CL
7.656.3%), followed by the group with a positive patch
testing and a negative medical history (41.3%, 95% CL
24.258.0%); the survivorship rate of the patients who had no
evidence of sensitization was 51.5% (95% CL 39.0%
63.3%). The statistical comparison of survival curves showed
signicant differences (p 0.05)
TABLE IV. Frequency (%) of Skin Positive Reactions to the
Bone Cement Components in Control Group and in Patients
with Cemented THR
Cement hapten
Group A
(n 66)
Group B
(n 14)
Group C
(n 50)
At least one hapten 9.1 9.6 11.8
Methyl-methacrylate 4.5 9.1 7.7
Butyl-methacrylate 1.5 0.0 0.0
Triethylene glycol dimethacrylate 2.0 0.0 0.0
Ethylene glycol dimethacrylate 3.0 0.0 3.8
N,N-Dimethyl-paratoluidine 1.5 0.0 3.8
Hydroxy-ethyl-methacrylate 1.5 0.0 3.8
Benzoyl-peroxide 3.0 0.0 0.0
Hydroquinone monobenzyl ether 3.0 0.0 7.7
TABLE V. Measures of Test Performance
a
Sensitivity Specicity
Likelihood Ratio
Positive Test
Likelihood Ratio
Negative Test
Patch testing result and medical history of metal allergy
b
At least one metal hapten 0.94 (0.861.02) 0.67 (0.600.74) 2.83 (2.273.52) 0.09 (0.020.34)
Patch testing result and outcome of THR
c
At least one hapten 0.43 (0.330.52) 0.45 (0.370.52) 0.83 (0.591.17) 1.17 (0.851.62)
Metal haptens 0.39 (0.290.487) 0.49 (0.360.63) 0.75 (0.531.08) 1.25 (0.921.71)
Bone-cement haptens 0.11 (0.050.18) 0.90 (0.820.98) 1.22 (0.453.28) 0.98 (0.8751.09)
a
Values in parentheses indicates 95% condence limits.
b
The number of allergic-positive test (True Positive, TP), the number of non allergic-positive test (False Positive, FP), the number of allergic-negative test (False Negative,
FN), and number non allergic-negative test (True Negative, TN) were placed in a two-by-two cross table.
c
The number of failed THR-positive test (TP), the number of stable THR-positive test (FP), the number of failed THR-negative test (FN), and number of stable
THR-negative test (TN) were placed in a two-by-two cross table. Sensitivity, TP/(TP FN); Specicity, TN/(TN FP); Likelihood Ratio Positive Test, Sensitivity/(1-Specicity);
and Likelihood Ratio Negative Test, (1-Sensitivity)/Specicity.
261 SENSITIZATION TO IMPLANT COMPONENTS
DISCUSSION
In spite of the preclinical evaluation of the medical device
safety,
2
hypersensitivity is an unpredictable phenomenon
that may occur in genetically predisposed patients after a
chronic exposure to immunogens and without a dose-
response relationship.
3
Therefore, some manufacturers of
orthopedic implants consider the metal sensitization
among the possible adverse effects, or suggest that when
material sensitivity is suspected, appropriate tests should
be made prior to selection of the device. In vitro tests,
based on the immune cell reactivity to metalprotein com-
plexes, have been addressed as reliable in diagnosing a
systemic metal-elicited hypersensitivity,
1012,14
but they
are poorly suitable for a large scale application because
complex, feasible only in qualied laboratories, and ex-
pensive in terms of time, reagents, and biological material.
The skin testing could overcome some of these limitations.
Although patch testing is considered as the gold standard
in diagnosing of contact allergy, some concerns exist about
its validity in detecting a systemic hypersensitivity, includ-
ing the immune reactivity in dermal contact, which is
likely different from that in the peri-implant tissues, as
well as the possibility of inducing hypersensitivity in a
previously insensitive patient.
14
Nevertheless, the assump-
tion that sensitization occurs in genetically predisposed
subjects decreases the clinical importance of both the
above remarks. In the former case, since the positive skin
reaction reects the susceptibility to the sensitization, it is
reasonable to assume that the immunopathological response
is irrespective of the contact site. In the latter one, if patch
testing is able to sensitize the patient, the continuous release
of degradation products from the implant could do the same.
The usefulness of patch testing in detecting the sensitiza-
tion to implant materials may be improved if patients are
tested with chemicals relevant to their problem.
19
The rst
step of the current study was to choose a panel of represen-
tative haptens that included all the components used in or-
thopedic implants. The choice of haptens and their concen-
tration was based on previous studies reporting that special
Figure 1. Kaplan-Meier survivorship analysis at 10-year end point. The statistical comparison of
survival curves was performed by the logrank Mantel-Cox test. Survivorship rates and 95% CL are
referred to in the text. (a) Eighty-three patients had negative () patch testing (PT), and 72 had positive
() PT; (b) 88 patients had negative () PT, 67 had positive () PT; (c) 135 patients had negative ()
PT, 20 had positive () PT; (d) 83 patients had no skin reaction, 53 patients had reaction to metal, 14
patients had reaction to metal and bone cement, and 5 patients had reaction to bone cement; and (e)
83 patients had negative PT and negative medical history for DTH (MH), 52 patients had positive PT
and negative MH, 19 patients had positive PT and positive MH.
262 GRANCHI ET AL.
preparations, such as vanadium and molybdenum, should be
used in concentration not over 2% because may produce an
irritating effect.
7
Generally, our results showed that sensitization to metals
was more frequent than that observed for bone cement, and it
was a common nding in patients candidates to THR. In other
hands, the incidence of skin allergy in the whole population
was found to be lower,
20
but other authors demonstrated that
when more haptens simultaneously are tested, the probability
of having a positive patch test increases.
21
For the most part,
patients included in pre-surgery group suffer a idiopathic
osteoarthritis of the hip, where a prevalence of Type-1 lym-
phocytes, the cell type associated with DTH, has been de-
scribed.
4,22
This immunological peculiarity could be a pre-
disposing factor to the sensitization and could explain the
high incidence of skin reactions that has been recorded in
candidates to THR. In our clinical series, nickel was the most
common metal sensitizer, followed by cobalt, chromium, and
manganese, while sporadic responses to vanadium were
found.
The frequency of positive patch testing to at least one
metal hapten in patients with THR did not differ from that
observed in patients before the implant, but signicant results
were found when the incidence was related to the metal
composition of the implant. According to the ndings of
other authors,
7
a high incidence of positive skin reaction to
vanadium was observed in patients who had a TiAlV com-
ponent. In patients with CoCrMo-failed implant, the remark-
able result was the lower frequency of metal DTH, namely
nickel hypersensitivity. Because nickel is the most common
metal sensitizer and its amount in both CoCrMo and TiAlV
alloys is very low, we may consider the incidence of nickel
sensitivity as informative for the immune status of the exam-
ined group. These results agree the conclusion of previous
studies, where the immunocompromised status of patients
who had a CoCrMo had been shown and related to the high
serum level of chromium and cobalt.
4,23,24
The type and the concentration of the haptens that have
been included in patch testing showed a good diagnostic
accuracy. The high sensitivity and the negative likelihood
ratio less than 0.1 indicate the excellent ability of the assay to
detect a metal allergy when it was previously documented;
18
on the contrary, the unsatisfactory values of specicity and
positive likelihood indicate that a large number of patients
with a negative medical history of skin allergy have a positive
patch test. This result conrms that the presence of the
implant inuences the incidence of sensitization, irrespective
of the clinical expression of contact dermatitis. Specicity
and positive likelihood ratio ameliorate when only positive
reactions with 2/3 grading were considered (data not
shown). Nevertheless the 1 reactions cannot be disregard-
ed: this type of reactivity is classied as possible contact
allergy, and some caution should be exercised in interpreting
the results,
16
but anyway it reect the capability of develop-
ing a response to the antigen challenge.
Our results partially agree with the ndings of other Au-
thors who used the most popular in vitro method, namely the
Lymphocyte Transformation Testing. In a recent work Hallab
et al.
14
showed that the incidence of sensitization to nickel,
cobalt, chromium, and titanium was 50% in preimplant pa-
tients, and the majority had a lymphocyte reactivity to nickel.
Like us, these authors found that the frequency of positive
responses to at least one metal did not change after THR,
except for the decrease in the incidence of sensitization to
nickel. The use of patch testing seems to add some benets:
more haptens may be tested, and more individuals who had a
medical history of contact allergy may be detected.
With regards to the status of the implant, no predictive
value may be ascribed to the presence of sensitization, as
likelihood ratios show that patch testing is not able to dis-
criminate between stable or failed THR. The informative
power of the patch testing emerges when the THR-survival
analysis is performed. The presence of sensitivity to at least
one hapten, as well as the medical history of DTH, inuences
negatively the outcome of the implant, because the THR
failure occurs earlier. When sensitivity to bone cement is
recorded, a swift decrease in the survival curve is observed,
and none of the patients has a stable THR at 10-year end
point.
In summary, our results show that patch testing may be a
suitable method for the large-scale detection of the sensitiza-
tion to the implant components, which is considered as an
unpredictable and adverse event. The usefulness of patch
testing does not exclude the reliability of the in vitro methods
that are advisable in doubtful cases, when the reactivity to
few haptens has to be conrmed. Finally, the current study
does not answer the question whether hypersensitivity is
cause or consequence of the implant failure: the hyper-
sensitivity may precede the prosthetic loosening, as well as
the generation of a large number of debris following the
motion of unstable implants may induce the sensitization of
genetically predisposed individuals. Although the causeef-
fect relationship between hypersensitivity and THR outcome
cannot be demonstrated, the hip prosthesis had a shorter
lifespan in patients who had a positive patch testing. These
ndings suggest that, irrespective to the onset time, sensiti-
zation may contribute signicantly to the initiation or perpet-
uation of the events that lead to the implant failure.
The authors wish to acknowledge Dr. Elettra Pignotti for her
assistance in the statistical management of the data, and Mr. Keith
Smith for his assistance in the editorial management of the manu-
script.
REFERENCES
1. Pizzoferrato A, Cenni E, Ciapetti G, Granchi D, Savarino L,
Stea S. Inammatory response to metals and ceramics. In:
Barbucci R, editor. Integrated Biomaterials Science. New York:
Kluwer Academic/Plenum Publishers; 2002. p 735791.
2. Food and Drug Administration (FDA).Center for Devices and
Radiological Health. Guidance for Industry and FDA Review-
ers. Immunotoxicity Testing Guidance, May 6, 1999. p 116.
3. Descotes J. Importance of immunotoxicity in safety assessment:
A medical toxicologists perspective. Toxicol Lett 2004;149:
103108.
263 SENSITIZATION TO IMPLANT COMPONENTS
4. Granchi D, Savarino L, Ciapetti G, Cenni E, Rotini R, Mieti M,
Baldini N, Giunti A. Immunological changes in patients with
primary osteoarthritis of the hip after total joint replacement.
J Bone Joint Surg Br 2003;85:758764.
5. Hallab N, Merritt K, Jacobs JJ. Metal sensitivity in patients with
orthopaedic implants. J Bone Joint Surg Am 2001;83:428436.
6. Lalor PA, Revell PA, Gray AB, Wright S, Railton GT, Freeman
MA. Sensitivity to titanium. A cause of implant failure. J Bone
Joint Surg Br 1991;73:2528.
7. Cancilleri F, De Giorgis P, Verdoia C, Parrini L, Lodi A, Crosti
C. Allergy to components of total hip arthroplasty before and
after surgery. Ital J Orthop Traumatol 1992;18:407410.
8. Gil-Albarova J, Lacleriga A, Barrios C, Canadell J. Lymphocyte
response to polymethylmethacrylate in loose total hip prosthe-
ses. J Bone Joint Surg Br 1992;74:825830.
9. Haddad FS, Cobb AG, Bentley G, Levell NJ, Dowd PM. Hy-
persensitivity in aseptic loosening of total hip replacements. The
role of constituents of bone cement. J Bone Joint Surg Br
1996;78:546549.
10. Hallab NJ, Mikecz K, Jacobs JJ. A triple assay technique for the
evaluation of metal-induced, delayed-type hypersensitivity re-
sponses in patients with or receiving total joint arthroplasty.
J Biomed Mater Res 2000;53:480489.
11. Granchi D, Ciapetti G, Savarino L, Stea S, Filippini F, Sudanese
A, Rotini R, Giunti A. Expression of the CD69 activation
antigen on lymphocytes of patients with hip prosthesis. Bioma-
terials 2000;21:20592065.
12. Valentine-Thon E, Schiwara HW. Validity of MELISA for
metal sensitivity testing. Neuro Endocrinol Lett 2003;24:5764.
13. Granchi D, Ciapetti G, Savarino L, Cavedagna D, Donati ME,
Pizzoferrato A. Assessment of metal extract toxicity on human
lymphocytes cultured in vitro. J Biomed Mater Res 1996;31:
183191.
14. Hallab NJ, Anderson S, Stafford T, Glant T, Jacobs JJ. Lym-
phocyte responses in patients with total hip arthroplasty. J Or-
thop Res 2005;23:384391.
15. Testoni M, Baruffaldi F, Mattioli P, Sudanese A, Terzi S,
Barbanti-Brodano G, Toni A. Evaluation of radiolucency con-
dition in total hip arthroplasty: A statistical comparison of the
diagnostic capability of digitised image vs. conventional X-ray
lm. Eur Radiol 2000;10:601608.
16. Drake LA, Dorner W, Goltz RW, Graham GF, Lewis CW,
Pariser DM, Salasche SJ, Skouge JW, Turner ML, Lowery BJ.
Guidelines of care for contact dermatitis. Committee on Guide-
lines of Care. J Am Acad Dermatol 1995;32:109113.
17. Bossuyt PM, Reitsma JB, Bruns DE, Gatsonis CA, Glasziou PP,
Irwig LM, Lijmer JG, Moher D, Rennie D, de Vet HC. Stan-
dards for reporting of diagnostic accuracy steering group. To-
wards complete and accurate reporting of studies of diagnostic
accuracy: The STARD initiative. BMJ 2003;326:4144.
18. Kocher MS, Zurakowski D. Clinical epidemiology and biosta-
tistics: A primer for orthopaedic surgeons. J Bone Joint Surg
Am 2004;86:607620.
19. Van der Valk PGM, Devos SA, Coenraads PJ. Evidence-based
diagnosis in patch testing. Contact Dermatitis 2003;48:121125.
20. Basketter DA, Briatico-Vangosa G, Kaestner W, Lally C, Bon-
tick WJ. Nickel, cobalt and chromium in consumer products: A
role in allergic contact dermatitis? Contact Dermatitis 1993;28:
1525.
21. Gawkrodger DJ, Lewis FM, Shah M. Contact sensitivity to
nickel and other metals in jewelry reactors. J Am Acad Derma-
tol 2000;43:3136.
22. Sakkas LI, Scanzello C, Johanson N, Burkholder J, Mitra A,
Salgame P, Katsetos CD, Platsoucas CD. T cells and T-cell
cytokine transcripts in the synovial membrane in patients with
osteoarthritis. Clin Diagn Lab Immunol 1998;5:430437.
23. Granchi D, Verri E, Ciapetti G, Savarino L, Cenni E, Gori A,
Pizzoferrato A. Effects of chromium extract on cytokine release
by mononuclear cells. Biomaterials 1998;19:283291.
24. Savarino L, Granchi D, Ciapetti G, Stea S, Donati ME, Zinghi
G, Fontanesi G, Rotini R, Montanaro L. Effects of metal ions on
white blood cells of patients with failed total joint arthroplasties.
J Biomed Mater Res 1999;47:543550.
264 GRANCHI ET AL.