1 Department of Clinical Chemistry and 2 Department of Obstetrics and Gynaecology, Trondheim University Hospital, Norway
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
Key words: analytical interference/heterophilic antibodies/immunoassay/monoclonal gammopathy/oestradiol
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
Case report |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
In the IVF treatment, pituitary suppression was obtained by daily 800 µg nafareline nasal spray (Synarela, Searle, Morpeth, Northumberland, UK) from day 19 in the previous cycle and for 16 days before starting ovarian stimulation with recombinant follicle stimulating hormone (FSH) 150 IU/day (Gonal-F, Ares-Serono, Geneva, Switzerland). She was monitored on day 10, after 9 days of FSH treatment. She appeared healthy, with no symptoms of OHSS. The ovaries measured 52x40 mm and 50x44 mm, with a total of two follicles of 18 mm, four of 15 mm, two of 14 mm and ~10 follicles <12 mm in diameter. The dose of FSH was reduced to 100 IU, and the serum oestradiol measurement was 56.50 nmol/l. From experience, we expected oestradiol concentrations to be <10 nmol/l. The treatment was cancelled, and the patient continued with the GnRH agonist. Daily monitoring for the following week showed the largest size of the ovaries to be 62x54 mm and 64x54 mm, and the oestradiol concentrations were 66.67, 52.05, 43.33, 20.19, 17.07 and 19.40 nmol/l. The patient never had any signs of OHSS, which would have been expected with these high oestradiol concentrations. The disagreement between the clinical and the biochemical observations led us to suspect a laboratory error or an oestradiol-producing tumour (Rey et al., 1996; Aboud, 1997
). Therefore, on two separate occasions oestradiol measurements were performed elsewhere using different methodology. The comparison of samples drawn on the same day, showed our laboratory and the reference laboratory values to be 20.19/0.19 nmol/l and 19.40/0.17 nmol/l respectively. The latter values of oestradiol were in concordance with the clinical situation, and this resulted in further examinations. Both oestradiol methods were automatic immunoassays utilizing polyclonal rabbit antibodies: ours was Elecsys Immunoassay (Boehringer Mannheim, Mannheim, Germany) and in the reference laboratory the method was AutoDelFIA (LKB Wallac, Turku, Finland). The Elecsys Immunoassay is an automated immunochemical electrochemiluminescence method where antibodyantigen reaction coupled with ruthenium chelate complex occurs on the surface of paramagnetic microparticles. In this assay, biotinylated anti-oestradiol rabbit antibodies and ruthenium-labelled anti-oestradiol antibodies form a `sandwich'-type complex and subsequently after adding streptavidin-coated microparticles the complexes bind to solid phase with biotinstreptavidin. The bound and free labels are measured using a magnetic field. The AutoDelFIA is an automated immunochemical immunofluorescence method which is also based on a sandwich-type assay utilizing anti-oestradiol rabbit antibodies and europium-labelled fluorescent complexes in a solid phase. The only major difference between these two methods is that AutoDelFIA does not work at saturating conditions regarding the solid phase antibody. Furthermore, all possible preclinical and methodological errors were excluded by a careful analysis.
The difference in oestradiol concentrations between these two methods was thought to be due to the presence of circulating heterophilic antibodies. These may have interfered with immunoassays. Therefore, further analyses were made from the blood sample showing 56.50 nmol/l, to neutralize possible circulating antibodies. The results were: 32.9 nmol/l using mouse serum as diluent and 22.3 nmol/l using rabbit serum respectively. It was obvious that some of the possible heterophilic antibodies were partly neutralized by rabbit serum. To characterize the pattern of possible anti-idiotypic type anti-rabbit response and especially because the result in the reference laboratory was normal, serum electrophoresis was performed. This investigation revealed a clear band in the gamma region. This was further confirmed by immunofixation, and an IgG-kappa type M-component was observed. This monoclonal fraction (15.5 g/l) disturbed our analyses, but not those of the reference laboratory. Subsequently, immediate extra blood samples were performed with the following results: leukocytes 7.7x109/l, haemoglobin 132 g/l, thrombocytes 216x109/l, reticulocytes 0.9%; differentiation count: neutrophilic 42%, lymphocytic 44%, monocytic 5%, eosinophilic 6%, basophilic 1 and 3% large unstained cells. Activated prothrombin time was 113% (normal range, 70130%). Serum concentrations were normal as follows: creatinine 78 µmol/l (55100 µmol/l), calcium 2.23 mmol/l (2.202.60 mmol/l), alanine aminotransferase 16 IU/l (<35 IU/l), gamma-glutananyl transferase 19 IU/l (<50 IU/l), haptoglobin 1.3 g/l (0.41.9 g/l), ß-2-microglobulin 0.9 mg/l (0.73.4 mg/l). The immunoglobulin concentrations were IgA 1.4 g/l (0.73.7 g/l), IgG 18.4 g/l (6.915.7 g/l), and IgM 1.4 g/l (0.62.3 g/l).
These data made us suspect that this patient might have had a haematological malignancy, most probably of lymphoid origin. The patient was referred to a haematologist, and bone marrow biopsy turned out to be normal with a normal plasma cell count. The diagnosis was consistent with benign monoclonal gammopathy.
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
Artefactually increased heterophilic antibodies in the patient's serum with specificity for immunoglobulin class G-kappa were found. These heterophilic antibodies formed complexes with the antisera to human oestradiol that are used in the enzyme-linked immunosorbent assay (ELISA) system, in particular blocking the binding of, but also partly blocking interaction with the second antibody. When it was measured immunometrically with sera of rabbit origin, the concentrations measured were significantly decreased. Protein electrophoresis was done to characterize the possible amount and immunoglobin subclass of circulating protein. Then an M-component was found and it was further characterized by immunofixation.
In large multicentre evaluation of two new enzyme-linked immunosorbent assays for FSH and LH, no interference by heterophilic antibodies was found (Thijssen et al., 1991). These methods are similar to our oestradiol analysis. These methods are known to be reproducible (coefficient of variation <5%), highly specific, and sensitive enough to measure the hormones directly in almost all patients' samples. However, many endogenous antibodies exhibit a potential for interference with immunoassays (Kohse and Wisser, 1990
). Interference by heterophilic antibodies can be abolished by the addition of non-immune serum. Investigations with our samples revealed an M-component cross-reacting with some polyclonal rabbit antibodies. This is possible, because the methods used for oestradiol determination used different polyclonal anti-rabbit antibodies. The possibility of anti-streptavidin antibodies was excluded using a binding assay. In clinical chemical analysis, autoantibodies, antibodies to foreign antigens, antibodies administered for therapeutic purposes, and monoclonal gammopathies are possible sources of interference. Differences between various immunometric assays and radioimmunoassays and their vulnerability to interference from heterophilic antibodies have been reported (Seth et al., 1989
). It is known from the laboratory experience of various manufacturers, although not reported, that heterophilic antibodies seldom disturb oestradiol analyses. No specific analyses of oestradiol assays have been reported, but we believe that this sort of disturbance must be very rare (perhaps <1/10 000 analyses) (Thijssen et al., 1991
; Seth et al., 1999). No specific data have been reported regarding monoclonal gammopathy as a source of laboratory errors. We know theoretically that it is possible, extremely rare, and that the prevalence is much lower than that of heterophilic antibodies.
There are various options available to avoid interference by heterophilic antibodies; alternatively, the sample can be studied elsewhere. The immunoassay itself has to be validated, as we have reported, and which led us to conclude that our analysis could not give the right result because of the clonality of the M-component. The sample can be preincubated with serum from other species, but this did not give an adequate result in our case. A separation procedure other than the double-antibody method, such as ether or polyethylene glycol extraction, may give better results (Dericks-Tan et al., 1984). However, all these techniques are very time consuming for automated routine laboratory analyses and require manual laboratory methods. Additionally, the respective hormone can always be measured in urine.
This case demonstrates how important it is to identify the nature of heterophilic antibodies, to delineate the processes that produce them, to examine the mechanisms by which these antibodies cause interference, and to explore how this information can be used to reduce immunoassay interference. This procedure may reveal another disease without any clinical significance so far, as it did in our case. Interference of this type has not been previously reported and its incidence is low, but also poorly characterized, because routine hormone laboratories do not often perform electrophoresis analyses. In fact, virtually nothing has been reported about monoclonal gammopathy as a source of laboratory errors. In our case the treatment of the patient was unnecessarily cancelled but on the other hand the procedure could have revealed a haematological cancer.
![]() |
Acknowledgments |
---|
![]() |
Notes |
---|
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
Dericks-Tan, J.S., Jost, A., Schwedes, U. and Taubert, H.D. (1984) Pseudohypergonadotropinemia and pseudohyperprolactinemia induced by heterophilic antibodies? Klin. Wochenschr, 62, 265273.[ISI][Medline]
Kohse, K.P. and Wisser, H. (1990) Antibodies as a source of analytical errors. J. Clin. Chem. Clin. Biochem., 28, 881892.[ISI][Medline]
Rey, R.A., L'Homme, C., Marcillac, I. et al. (1996) Antimüllerian hormone as a serum marker of granulosa cell tumors of the ovary: comparative study with serum alpha-inhibin and oestradiol. Am. J. Obstet. Gynecol., 174, 958965.[ISI][Medline]
Seth, J., Hanning, I., Bacon, R.R. and Hunter, W.M. (1989) Progress and problems in immunoassays for serum pituitary gonadotrophins: evidence from the UK external quality assessment schemes, (EQAS) 19801988. Clin. Chim. Acta, 186, 6782.[ISI][Medline]
Thijssen, J.H., Wood, W.G., Kessler, A.C. et al. (1991) Multicenter evaluation of new enzyme-linked immunoassays of follitropin and lutropin in serum or plasma. Clin. Chem., 37, 12571263.
Waldenstrøm, U., Kahn, J.A., Marsk, L. and Nilsson, S. (1999) High pregnancy rates and successful prevention of severe ovarian hyperstimulation syndrome by `prolonged coasting' of hyperstimulated patients: a multicentre study. Hum. Reprod., 14, 294297.
Submitted on April 23, 1999; accepted on July 27, 1999.