Nickel hypersensitivity and skin patch testing in total hip replacement surgery: a systematic review

Search strategy and study selection

Systematic electronic literature searches were conducted in PubMed, Ovid and Healthcare Database Advanced Search (HDAS) searching EMBASE and Medline databases (until 13 April 2021). Combinations of medical subject heading (MeSH) terms and keywords were used to identify relevant papers with a high level of sensitivity. Table 1 shows the search string applied in the search. Further manual searches of the reference lists of the papers and searching the grey literature supplemented the systematic electronic search. Papers were screened initially by title and abstract. Two independent reviewers screened the selected studies and the results of the search strategy were reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) tool.

Eligibility criteria

The inclusion criteria were: (1) clinical studies on testing nickel hypersensitivity with patch testing in patients with THR; (2) published in English or with translation freely available; (3) full text of studies available. Exclusion criteria included (1) case reports; (2) review studies; (3) cadaver studies and (4) no reported outcome.

Data extraction/analysis

The level of evidence (LE) was assessed based on previously published criteria from the Oxford Centre for Evidence-Based Medicine¹³ and the methodological quality was assessed using the Methodological Index for Non-Randomised Studies (MINORS) tool.¹⁴ The following information was obtained from each study:

1. Study characteristics (e.g. author, geographical area, study design)

2. Patient characteristics (e.g. number of patients included, and number of hip joints operated on, age, gender)

3. Implant characteristics (e.g. type of implant, bearing)

4. Details on patch testing (patch substances used; time point at which patch test was performed)

5. Prevalence of nickel hypersensitivity (before and/or after surgery)

6. Clinical results (e.g. complications, stable or failed implant, adverse reaction to metal debris, systemic adverse reactions)

7. Main conclusions and recommendations

Search results

Our initial literature search using the MeSH terms detailed in Table 1 identified 795 studies, and after the duplicates were removed 439 remained. The abstracts of these papers were screened. Twenty-six clinical studies met the inclusion criteria for this review (Fig. 1).

View Full Size

Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) study selection flow diagram.

Citation: EFORT Open Reviews 6, 10; 10.1302/2058-5241.6.210051

• Download Figure
• Download figure as PowerPoint slide

Quality assessment

All studies were of small to medium size and focused on metal and/or nickel sensitivity in patients who underwent total hip arthroplasty. Seven of the studies^{18,20,24,26–28,39} had a cohort study design with LE of III, while the other 19 were case-series^{16,21,25,29–34,37,40} or case-controlled^{15,17,19,22,23,35,36,38} with LE of IV. The average MINORS score was 12.2 (Table 2).

Cohort characteristics

Across all of the studies there were 3466 participants with an average age of 63.5 years (range 48–71) and average female proportion of 65.7% (range 24–89%). A total of 1852 primary and revision hip arthroplasties were performed, either prior to the conducted studies or as part of them. The rest of the participants either comprised control groups or underwent a different procedure, such as total knee replacement (TKR), or open reduction and internal fixation (ORIF). A detailed description of the study populations is shown in Table 3.

Implant characteristics

The implant type and type of bearing was documented in 15 studies,^{15–17,19–23,26,27,30–33,39} with only the implant type being recorded in four studies^18,24,35,36 and only the type of bearing in a further four.^25,37,38,40 In the three remaining studies^28,29,34 there was no clear documentation of either the type of implant used or the bearing. A breakdown of the type of implants used and the bearing from each study is shown in Table 3.

Patch testing

All 26 studies used patch testing to identify metal hypersensitivity. The substances applied in the patch testing along with the exact concentration of each substance were listed in 20 studies (Table 4). Thirteen of the studies^{15,17,18,21–23,25,27–30,34,38} used 5% nickel sulphate, four ^20,31,33,39 used 2.5% nickel sulphate, and the remaining three used nickel sulphate in concentrations of 1%,¹⁹ 2%¹⁶ and 3%.³² In the other six studies^{24,26,35–37,40} the strength of the nickel sulphate used was not documented.

In ten studies^{17,18,20,22,28,31,33,34,36,39} the participants were patch tested both before and after their operation, and in 14 studies^{15,16,19,21,23–27,29,30,32,35,37} participants were patch tested only after the primary arthroplasty took place. In three studies with patients undergoing revision THR, the timing of patch testing was not documented in relation to the revision procedure.^19,30,37 Additionally, in one study six patients had patch testing after the revision; however, for the remaining patients it was not documented when patch testing was performed.¹⁶ In the study by Thyssen et al, patch testing was performed prior to THR in 292 cases (82%) and in 64 cases (18%) after THR was performed.³⁸ In one study patch testing was performed prior to primary THR.⁴⁰

The reported time until patch testing was performed post-operatively ranged from three months to 18 years (Table 5).

Prevalence of nickel sensitivity

The prevalence of nickel sensitivity in each study ranged from 1.5% to 33.3%, and two studies^16,32 reported no positive reaction to nickel amongst the participants tested. Two studies^18,24 had selected only participants with a known hypersensitivity to nickel and subsequently reported nickel hypersensitivity on patch testing of 76.5%²⁴ and 83.3%¹⁸ (Table 5, highlighted in pink).

Ten studies^{17,18,20,22,28,31,33,34,36,39} compared the pre-operative and post-operative prevalence of nickel hypersensitivity in the same groups of patients and in five of these^{17,28,31,34,39} it was reported there was an increase in the number of patients testing positive for nickel hypersensitivity post-operatively. Kręcisz et al concluded that the increase in nickel hypersensitivity prevalence was minimal (from 20% pre-operatively to 20.8% post-operatively), but it was noted that three patients had developed a new hypersensitivity to nickel following surgery.²⁸

Four studies^18,20,33,36 noted a decrease in the number of patients hypersensitive to nickel after the operation. One study²² did not report the results of patch tests post-operatively.

Study recommendations

Eight studies^{20–22,24,28,31,34,39} concluded that orthopaedic implants could trigger metal hypersensitivity in patients, but that the relationship between the hypersensitivity and subsequent implant failure or loosening remained unknown.^21,28 Three studies reported a relationship between metal hypersensitivity and prosthesis loosening,¹⁵ higher patch test reactivity in arthroplasty patients experiencing complications,³⁵ and a correlation between metal hypersensitivity and post-surgical thigh pain.⁴⁰

Seven of the studies^{16,18,19,29,30,33,38} concluded that they did not support a possible relationship between metal hypersensitivity and THR complications, implant loosening or the need for revision. Two of these studies^18,33 and one further study³² concluded that the release of metal ions did not result in increased hypersensitivity.

Gustafson et al reported that, despite metal ion concentrations being higher in patients with metal-on-metal bearings, compared to those with metal-on-plastic articulations, there was no difference in the prevalence of metal hypersensitivity between the two groups²⁵ with Hjorth et al reporting no association between the formation of pseudotumours and serum metal-ion levels, metal patch test reactivity or atopic dermatitis in patients with metal-on-metal bearings.²⁷ Two studies investigated lymphocyte-mediated hyperactivity to metals rather than patch test reactions, but the clinical implications of such hyperactivity in patients with THR remained unknown.^26,37 One study concluded that it was doubtful that metal hypersensitivity was triggered by THR.¹⁷ Granchi et al reported that it had not been possible to establish a cause–effect relationship between sensitization and THR complications but, reported a shorter THR lifespan in patients with a positive result to patch testing.²³ One study did not comment on the possibility of sensitization or any potential relationship between metal hypersensitivity and THR complications.³⁶

Twelve studies concluded that patch testing was a valuable tool,^{16,20,22–25,28,31,34–36,40} one of which recommended that it should be mandatory,²⁸ with three further studies recommending its targeted use.^24,25,34 Five of these studies concluded that patch testing was a valuable diagnostic tool in the detection of metal hypersensitivity,^16,35,40 even when the testing was delayed³⁶ and that testing might have an application in identifying sensitization to implants on a larger scale.²³ Four studies^20,22,28,31 recommended routine patch testing for all patients pre-operatively, one of which suggested that testing should be obligatory.²⁸ Three studies recommended considering the clinical relevance of patch tests²⁵ and only to perform this investigation when there was a known history of hypersensitivity reactions.^24,34

Three studies reported that routine patch testing was not required^33,39 or that it was unrealistic.¹⁷ One study concluded that patch testing was a poor diagnostic tool and might not be sufficient to accurately demonstrate an adaptive immune response.²⁶

Ten of the studies did not comment on the usefulness of patch testing in identifying nickel hypersensitivities (Table 6).^{15,18,19,21,27,29,30,32,37,38}

Patch testing

The systematic review confirmed that there was no consensus on the routine use of patch testing, but the studies were generally consistent in the chemical constituents that were used for the patch testing, although there was a wide range in the timing of administration. Some studies suggested that patch testing was a reliable, gold standard tool in establishing nickel hypersensitivity^{16,23,35,36,40} and that it should even be mandatory,²⁸ but a similar number recommended that patch testing was not routinely required.^17,33,39

A study by Thomas et al evaluated the usefulness of late reading of the patch testing. It reported an overall positive reaction to nickel in 32 patients (12.8%). Eleven of those positive reactions (34.4%) were recorded following a late reading of the patch test at day 6.³⁶ Reed et al, evaluating the usefulness of patch testing in the guidance of implant choice, concluded that patch testing might be helpful prior to operation, but had limited value post-operatively.⁴¹ Furthermore, Hallab et al reported that patch testing was a poor diagnostic tool and suggested that there was no correlation with ion levels or measures of hypersensitivity and that there was no correlation with potential adaptive immune response in the deep tissue.²⁶

There is evidence that patch tests have high sensitivity and specificity to detect hypersensitivity, but the immunologic response which occurs is triggered by the intradermal Langerhans cells, whereas the metal hypersensitivity reaction in the joint space is mediated by different mechanisms involving macrophages and lymphocytes.⁵ Christiansen et al reported that there was a positive correlation between failure of joint arthroplasty and metal hypersensitivity, investigated by in vitro assay on peripheral blood lymphocytes, and that the findings were suggestive that prosthesis failure could be attributed to a cell-mediated immunity to metals.⁴²

It is not therefore clear whether patch testing can accurately predict outcomes and complications following THR.⁵ Lhotka et al reported a possible relationship between nickel hypersensitivity and reactions to metallic skin clips used for wound closure, but none of the studies included in this review specifically commented on this issue.⁴³

Nickel hypersensitivity prevalence

It has been reported that the prevalence of nickel sensitivity in the general population is approximately 13%,⁴ but the prevalence of nickel hypersensitivity following patch testing in the studies reviewed was reported to range from 1.5% to 33.3%. This discrepancy can be explained by the number of participants in each study, the inclusion and exclusion criteria, as well as the lack of uniform reporting of the nickel hypersensitivity. Eight of the studies supported the concept that THR triggers metal hypersensitivity in patients,^{20,21,22,24,28,31,34,39} but in four there was a decrease in nickel hypersensitivity prevalence post-surgery.^18,20,33,36 Possible explanations could be false positive results pre-operatively, or false negative results following surgery, or development of immunological tolerance.³³

Nineteen of the studies included patients who underwent primary THR.^{15,17,18,20–23,25–29,31–35,39,40} The nickel hypersensitivity prevalence ranged from 0.0–33.3% across 18 of those studies, while in one study the prevalence was 83.3% as per inclusion criteria.¹⁸ Five studies looked at patients awaiting revision THR.^{16,19,24,30,37} Four of those studies reported nickel hypersensitivity prevalence of 0.0% to 25.0%, whereas in one study patients undergoing revision THR had known nickel hypersensitivity and the prevalence was 76.5%.²⁴ The study by Thyssen et al looked at both primary and revision cases and reported nickel reaction in 11% of the patients with primary THR, 10% in patients undergoing one revision, and 0% in patients undergoing two or three revisions.³⁸ One study investigated metal hypersensitivity in patients with both primary and revision THR; however, it did not comment on the prevalence of each group separately.³⁶ Given the wide range of participants included in each study and the reported nickel hypersensitivity prevalence, it is impossible to compare the sensitivity rates between the two groups.

Implant type and bearing

A variety of implants and types of bearing were featured in the studies reviewed, but only 15 of the 26 studies^{15–17,19–23,26,27,30–33,39} clearly reported the details of the implant used as well as the bearing or a breakdown of number of patients. Three of the studies^28,29,34 reported neither and this made it impossible to compare the nickel hypersensitivity prevalence between patient groups with different implant types or bearings.

Davies et al investigated peri-prosthetic tissue samples from metal-on-metal (MOM) and metal-on-plastic (MOP) THR and compared them to control samples from patients undergoing primary hip replacement. They observed a distinct and different pattern and type of inflammation between the samples, reporting that MOM tissue samples had a more prominent ulcerated appearance with extensive lymphocytic infiltration, while MOP tissue samples were less ulcerated with no plasma cell or lymphocytic infiltration.⁴⁴ A study by Brien et al reported that loosening of titanium-alloy implants led to disproportionally high levels of titanium and vanadium in synovial fluid and surrounding tissues when compared to cobalt, chromium and nickel levels released from loosened cobalt-chromium or stainless steel implants.⁴⁵ Although they raised concerns about the metallosis that could occur, it was unclear what effect this had on the eventual outcome of the THR.⁴⁵

Limitations

There are several limitations in this systematic review, which include the low level of evidence of the studies, the limited number of patients involved in some of them, the methodological variability of the studies and the inadequate reporting of the results of certain studies. While the participant groups appeared similar across all of the studies, it was not possible to directly compare the prevalence of nickel hypersensitivity due to the lack of uniform reporting of the number of participants with positive patch tests in the THR and the control groups.

Several of the articles compared groups of patients undergoing not only hip but also knee^22,28,36,40 and shoulder³⁴ arthroplasties. However, the results of the patch testing of those patients were not stratified by the operation undergone, but only as a cohort.^{18,22,24,28,34–36,40} Eleven of the studies were published in the last 10 years^{19,22,24–28,34–36,40} but the review also included studies dating back to 1975, with 12 of the papers being published in 1997 or earlier.^{15–18,20,21,29–33,39} Despite these limitations, it was still possible to draw some conclusions.