MELISA critique

MELISA critique

The below articles critique the MELISA test. The first article asserts that MELISA gives false positive reactions and therefore cannot be used. This is untrue. The critique uses patch testing as a gold standard for the diagnosis of metal allergy but patch testing is not sensitive enough to identify all metal-allergic patients and therefore some patch test negative patients may be positive in the MELISA test.


In the second article, the authors claim that MELISA “cannot identify patients with systemic symptoms attributed to dental amalgam”. In this study, the criterion for patients to be included was the patient’s own suspicion that they were “mercury poisoned”. MELISA is a test for metal allergy, not for mercury poisoning, and certainly not for suspected mercury poisoning. In addition to this, patients with allergies and oral mucosal changes, who might have been allergic to mercury, were excluded from the study. Out of 150 referred patients, only 23 were included in the study. This is incompatible with good scientific practice.

For more details, please read the discussion below.


1. Cederbrant et al., In vitro lymphocyte proliferation as compared to patch test using gold, palladium and nickel, Int Arch Allergy Appl Immunol; 1997;112:212-217

2. Cederbrant et al., In vitro lymphoproliferative assays with HgCl2 cannot identify patients with systemic symptoms attributed to dental amalgam, J Dent Res., 1999; 78(8):1450-8


Critical evaluation of the article In vitro lymphocyte proliferation as compared to patch test using gold, palladium and nickel


The major problem with this article is that the sensitivity and specificity of in vitro lymphocyte proliferation tests were evaluated with the use of patch test as a reference. It is well known, that application of metal salts under the occlusion on the skin of the back, in a vehicle such as petroleatum, might often cause irritative (toxic) local reaction. Dermatologists often call such tests false-positive. This is mainly true for inorganic mercury and organic mercury salts but is quite common even for testing with other metal salts, such as gold salts. Originally, gold chloride has been used for patch testing but this has been stopped because of its irritative properties.


The second problem is that epicutaneous application of metal salts on the skin might not properly reflect the local oral situation, which is far more aggressive. Metal corrosion often occurs in the presence of bacteria and galvanic streams, generated by the presence of dental metals with different potentials and the conditions of metal antigen presentation may be considerably different to those in the skin.


Patch test may be a valuable tool for diagnosis of contact dermatitis, where sensitization occurs through the skin, but the clinical relevance of patch testing for the diagnosis of allergy due to dental restorative materials is doubtful.


It is well known that positive patch test to gold in the absence of mucosal changes around the gold restorations is generally deemed “not clinically relevant”.


Schaffran et al (17) determined the prevalence of positive patch tests to gold in asymptomatic individuals with gold restorations and compared the results with those obtained in healthy subjects without gold restorations. Out of 71 healthy subjects with dental gold, 24 (34%) had positive patch test to gold versus 7 out of 67 gold-free subjects (11%). Of those with positive gold reaction, 12 out of 31 (39%) also had a positive nickel reaction. Nickel alone was positive in 18 of 71 (25%) of gold-subjects versus 11 of 65 (17%) of non gold subjects. Nineteen out of 29 (66%) of those with a positive nickel reaction also reacted to palladium and 19 out of 22 (86%) of those with a palladium reaction also reacted to nickel. Thus, patch tests with metals frequently show “false positivity” and therefore cannot be used for evaluation of other diagnostic tests.


Having said that, let’s look at the actual results presented by Cederbrant et al. (1).


The specificity of an optimized lymphocyte proliferation test (MELISA) is relatively high: 79% for the detection of gold allergy, and 95% for the detection of palladium allergy.


Regarding sensitivity, the sensitivity of MELISA to detect palladium allergy is again surprisingly high: 95%. Sensitivity regarding detection of gold allergy is lower (55%) but this could be due to irritation rather than the allergic nature of local gold reactions.


Regarding nickel, many patients with known nickel allergy are patch test-negative, so the sensitivity of patch testing for the detection of nickel allergy is low. False-positive metal patch test results occur as well, especially if metal salts are tested close to each other. For example, a strong positive nickel reaction creates a non specific hypersensitivity of the surrounding skin and induces false-positive reactions to other metals (18).


Local irritative reactions have nothing to do with systemic metal allergy, which is measured at the level of sensitized memory cells in MELISA and under those conditions, the MELISA test is invariably negative.


The case of false over-representation of allergy to Kathon CG, as diagnosed by patch testing, has been published (19). Double-blind provocation (use test) with lotion containing Kathon CG, was positive in only half of the patients exhibiting positive patch test. The provocation-positive patients did have Kathon CG-specific lymphocytes in the blood thus showing the systemic allergy in only 50% of patch test-positive patients.


Critical evaluation of the article Cederbrant et al 1999 (2)

The major drawback of this study is the extreme selection of included patients. From 150 patients originally referred for evaluation for possible dental metal-induced side-effects, only 23 (15%) were included. The inclusion criteria were:


1. Alleged amalgam-related illness, characterized by unspecific multi-symptoms of psychological, sensory or neurological nature.
2. No established diseases and absence of local mucosal changes.
3. Presence of amalgam only in the oral cavity.
4. Age 30-60 years. (The mean age was not given).


It is obvious, that most patients suffering from side-effects from dental restorative materials might have local mucosal changes or burning and itching, and they often suffer from various autoimmune and chronic inflammatory diseases (20, 21).


The increased frequency of autoantibodies was noted by Lundström in 30% of patients with OLP but only in 9% of the controls (20). 55% of patients with autoimmune diseases such as primary Sjögren’s syndrome suffer from oral dysfunction and 18% from lichenoid lesions (21). Similar patients might all have been excluded from the study despite the fact that mercury has been implicated as a possible causative and acceleration factor of certain rheumatoid diseases (22, 23) and the test evaluated (MELISA) is based on the measurement of immunological reaction.


Most adult patients aged 30-60 also have other dental metal restorations, in addition to amalgams, such as gold crowns and metal bound ceramics. The authors tried to characterize the “amalgam group” by measuring the mercury output in the nightly urine and compare it to mercury concentration found in the urine of healthy amalgam controls. There was no difference in the urinary excretion in patients and controls. Hence, the authors present no evidence that the alleged patient group was really affected by amalgam. It is clear, that under such conditions, MELISA cannot identify any amalgam-sick patients: there might not be any!


The other serious problem in the article is the selection of control patients and evaluation of sensitivity and specificity of MELISA by using the patch test to inorganic mercury as a reference (gold standard). It is generally accepted that patch test performed with inorganic mercury in patients with oral lichen planus or other oral diseases is highly unreliable, due to irritative (toxic) effects of inorganic mercury salts on the skin. For example, Skoglund et al. (14) presented evidence of the healing of OLP around amalgam fillings after the replacement of amalgam fillings with a non-metallic alternative. The healing was observed in the majority of patients and was independent of the results of patch test. Thus, the clinical relevance of patch testing with inorganic mercury is limited and such a test cannot be used as a gold standard to evaluate another method.


The two control groups consisted of 30 healthy amalgam bearers and 10 amalgam-free controls. Patch testing was used to evaluate the “real” allergic status of controls and patients. The results were as follows: Six healthy subjects out of 27 tested (22%) and two amalgam-free controls out of 6 tested (33%) turned out to be positive in the patch test.


This should be compared to the results of patch testing of the patient group: out of 23 patients tested, 6 patients were patch test positive (26%). Thus, according to patch testing, there was no difference in the allergy status of patients and of controls. The control groups were biased by the presence of subjects susceptible to mercury (one amalgam-free subject was patch test positive to mercury and another one to nickel) or to other metals (5 healthy amalgam bearers showed positive patch test to nickel and one to cobalt and chromium).


Disregarding this, the specificity of MELISA testing, based on the in vitro testing of subjects with negative patch test, was 70%, about the same as for patch testing. The situation is even worse when the authors claim that the specificity of MELISA performed in 4 patch test-negative amalgam-free subjects is 60%. This calculation, based on 4 patients after the removal of two patch test-positive amalgam-free patients, is not based on proper science.


The conclusion that the study has not been set up properly right from the beginning is supported by findings in the patient questionnaires on allergy. Overall, there were no statistically significant differences in hereditary family allergy or in the presence of clinical metal allergy in patients compared to “controls”. Where the latter is concerned, 43% of patients reported allergy to nickel, compared to 18% of controls. Control “healthy” subjects reported allergic conjunctivitis (23%) and allergic rhinitis (15%). Those results were not statistically different from the patient group indicating that with regard to allergic status (susceptibility), the patient and controls were similar. Since MELISA measures immune reactivity, and this characteristic was the same for both the control and patient group, it is given that MELISA could not discriminate between patients and controls.


1. Cederbrant et al., In vitro lymphocyte proliferation as compared to patch test using gold, palladium and nickel, Int Arch Allergy Appl Immunol; 1997;112:212-217
2. Cederbrant et al., In vitro lymphoproliferative assays with HgCl2 cannot identify patients with systemic symptoms attributed to dental amalgam, J Dent Res., 1999; 78(8):1450-8

3. Stejskal et al., MELISA® – an in vitro tool for the study of metal allergy, Toxicology In Vitro, 1994:8; 991-1000

4. Stejskal et al., Metal-specific lymphocytes: biomarkers of sensitivity in man, Neuro Endocrinol Lett, 1999;20(5):289-298

5. Tibbling et al., Immunological and Brain MRI Changes in Patients with Suspected Metal Intoxication, International Journal of Occupational Medicine and Toxicology, 1995;4:2

6. Stejskal et al., Mercury-Specific Lymphocytes: An indication of Mercury Allergy in Man, Journal of Clinical Immunology, 1996;16:1

7. F Wantke et al., Patch test reactions in children, adults and the elderly, A comparative study in patients with suspected allergic contact dermatitis, Contact Dermatitis 1996; 34: 316-319

8. Geier D, Geier MR., Neurodevelopmental disorders following thimerosal-containing childhood immunizations: a follow-up analysis, Int J Toxicol., 2004;23:369-76

9. Valentine-Thon et al., Validity of MELISA® for metal sensitivity testing, Neuroendocrinology Letters, 2003;242:57-64

10. Laiendecker R, van Joost T, Oral manifestation of gold allergy, J Am Acad Dermatol, 1994:30:205-209

11. Issa Y et al., Healing of oral lichenoid lesions after replacing amalgam restorations: a systemic review, Oral Surg Med Oral Pathol Oral Radiol Endod, 2004;98:553-65

12. Bruce GJ, Hall WB, Nickel hypersensitivity-related periodontitis, Compend Contin Educ Dent, 1995;178:180-184

13. Tosti A et al., Contact and irritant stomatitis, Semin Cutan Med Surg, 1997;16:314-319

14. Skoglund A, Value of epicutaneous patch testing in patients with oral, mucosal lesions of lichenoid character, Scand J Dent. Res, 1994;102:216-222

15. Östman PO et al., Amalgam-associated oral lichenoid reactions, Clinical and histological changes after removal of amalgam fillings, Oral Surg Oral Med Oral Pathol Oral Radiol Endod, 1996;81:459-465

16. Bigazzi PL, Chang L, ed, Toxicology of metals; Lewis publishers, CRC press inc USA, 1996

17. Schaffran RM, et al., Prevalence of gold sensitivity in asymptomatic individuals with gold dental restorations. Am J Contact Dermat, 1999;10:201-206

18. RL Rietchel and JF Fowler Eds, Contact dermatitis and other reactions to metals,
Lippincott, Williams and Wilkins, Fishers contact dermatitis, p. 613

19. Stejskal et al., Lymphocyte Transformation Test for Diagnosis of Isothiazolinone
Allergy in Man, J Invest Dermatol 94:789-802, 1990

20. Lundström IM, Allergy and corrosion of dental materials in patients with oral lichen
planus, Int J Oral Surg., 1984;13(1):16-24

21. Lundström IM, Lindström FD, Subjective and clinical oral symptoms in patients with
primary Sjogren’s syndrome, Clin Exp Rheumatol., 1995;13(6):725-31
22. Cooper GS et al., Occupational risk factors for the development of systemic
lupus erythematosus, J Rheumatol., 2004;31(10):1928-33

23. Hudson et al., Susceptibility of lupus-prone NZM mouse strains to lead exacerbation
of systemic lupus erythematosus symptoms, Toxicol Environ Health A.,

Comments: When a publication has three authors or more, only the first author is mentioned together with ‘et al’.