SENSITIVE TEST FOR THYROID HORMONE AUTOANTIBODIES IN SERUM
TABLE II-PERCENTAGE OF
1-LABELLED TRACERS PRECIPITATED
SIR,-During routine assays of free thyroxine (FT4) by the ’Amerlex FT4 RIA’ kit (Amersham International) we encountered sera from four women which gave apparently conflicting results (table I). The very high FT4 values indicated severe hyperTABLE I-THYROID HORMONE STUDIES IN FOUR PATIENTS WITH ANOMALOUS
FT4 VALUES AND IN CONTROLS
thyroidism but total T4 and triiodothyronine (T3) values (Amerlex RIA) did not support these implications. Thyroid stimulating hormone (TSH) values were much increased in two cases, and the patients were judged to be clinically hypothyroid. These results hinted at interference in the FT4 assay, possibly from autoantibodies directed against T4. 1-5 Endogenous autoantibodies could bind the 125 I-labelled tracer used in the assay, removing tracer from the RIA reaction, and thus give artificially high values. This effect should be especially noticeable in the Amersham FT4 assay because the analogue used as a tracer is virtually incapable of binding with the normal serum protein.Avid thyroid hormone autoantibodies could thus profoundly increase the effective serum binding of the tracer.7,8 1251-labelled T4 or T3 should give much smaller effects, because these tracers will also be partly bound by serum proteins, even when blocking agent is present. Assays for total T4 and T3 should be less affected, as less serum is used and antibody and labelled tracer is added in the RIA reaction. To test this hypothesis, 100 1 serum was incubated at 37°C for 1 h with 0 - 5 ml of the 125 I-labelled tracer used in the Amerlex FT4 assay. 1 ml of 200/o (w/v) polyethyleneglycol (PEG) was added, and, after 15 min incubation at room temperature, the serum gammaglobulins were centrifuged and counted. Parallel experiments were done using l25I-labelled T4 or T3. Results from controls, with varying FT4 concentrations and thyroid function tests consistent with the clinical condition, were also included. Table 11 shows that, where anomalies in FT4 values were encountered in the direct RIA method, a significant proportion of the 125I-tracer was precipitated by PEG. In most cases, a higher proportion of the analogue was precipitated than with T4 or T3. Not all patients with autoantibodies bound T4, T3, or the T4 analogue equivalently. This may reflect different populations of antibodies, or differences in cross-reactivity for given antibody
is often more sensitive than when T4 or T3 are used. Levels of antibodies that may hitherto have gone undetected by previous methods may now be within detection range. As patients with serum antibodies directed against thyroidal components may be signalling present or future evidence of thyroidal malfunction, it is important to maximise the sensitivity of testing methodologies. The effect of marginally raised concentrations of such antibodies on the Amerlex free thyroid hormone assay is important only when the binding potential (concentration x binding constant) becomes significant compared with that of the very avid antiserum used in the assay process. Weak autoantibodies found in most of this rare group will therefore have negligible effects on the FT4 assay. Nevertheless, anomalously high free thyroid hormone estimates by RIA, caused by serum autoantibodies, can be readily rationalised by this simple technique. Department of Pathology, Mount Vernon Hospital, Northwood, Middx
Clinical Reagents Development Department, Amersham International plc, Amersham, Bucks
F. MURPHY C. A. NEEDHAM N. BARRON T. A. WILKINS J. E. M. MIDGLEY
Our results show that the technique of using the 1 251-labelled T4 analogue as a detector of autoantibodies in patients with thyroiditis J, Rall JE, Rawson RW. An unusual instance of thyroxine binding by human y-globulin. J Clin Endocrinol Metab 1956; 16: 573-79. 2. Premachandra BN, Blumenthal HT. Abnormal binding of thyroid hormone in sera from patients with Hashimoto’s disease. J Clin Endocrinol Metab 1967; 27: 931-36. 3. Staeheli V, Vallotton MB, Burger A. Detection of human antithyroxine and antitriiodothyronine antibodies in different thyroid conditions. J Clin Endocrinol Metab 1975; 41: 669—75. 4. Karlsson FA, Wibell L, Wide L. Hypothyroidism due to thyroid-hormone-binding antibodies. N Engl J Med 1977; 296: 1146-48. 5. Ginsberg J, Segal D, Ehrlich RM, Walfish PG. Inappropriate triiodothyronine (T3) and thyroxine (T4) radioimmunoassay levels secondary to circulating thyroid 1. Robbins
hormone autoantibodies. Clin Endocrinol 1978; 8: 133-39. 6. Wilkins TA, Midgley JEM. New methods of free thyroid hormone assay. In: Bizollon CA, ed. Physiological peptides and trends in radioimmunoassay Amsterdam: Elsevier, North Holland, 1981: 215-34. 7. Konishi J, Iida Y, Kousaka T, Ikekubo K, Nakagawa T, Torizuka K Effect of antithyroxine autoantibodies on radioimmunoassay of free thyroxine in serum. Clin Chem 1982; 28: 1389-91. 8. Mullinger RN, Walker G Free thyroxine in thyroid disease. Clin Chem 1982; 28:
MIS-LEAD BY THE E.E.C.
SIR,-The European Economic Community (E.E.C.) has issued directives on lead levels,1,2 which are to become mandatory. Have the E.E.C. commissioners experience of guidelines in other areas-for example, of "normal values" in haematology of recommended dietary allowances in nutrition? Or do they believe that all the issues to do with lead are more fully worked out than is the case with, say, iron added to flour or airborne dust in flaxpreparing rooms? If not, should they not have commissioned research to provide the missing knowledge before issuing directives which mean enormous expenditure for uncertain benefits? The directive on water for human consumption’ states that, where lead pipes are present, the lead content "should not exceed 50 fg/1 in a sample taken after flushing. If the sample is taken either directly or after flushing and the lead content either frequently or to an appreciable extent exceeds 100 g/1, suitable measures must be taken to reduce the exposure to lead on the part of the consumer". This directive contains so many value judgments ("frequently", "to an appreciable extent") that it is likely to be unworkable. Furthermore, "flushing" is a curious term to find in a mandatory directive: if thoroughly done, flushing can reduce a high lead-level to less than one-tenth so that almost any supply could be shown to conform. In water sampling the variability of repeated estimations is enormous, but the most reproducible are those made on "first flush" samples, drawn after overnight standing. These are difficult samples to obtain and common practice seems to be for "random daytime" sampling, some of which will approximate to "first flush" samples if the tap sampled has not been used for a few hours, while others will approximate to "running" if the housewife has been busy at the kitchen sink just before the sample is taken. "Random 1. Council of the 2. Council of the
European Communities. OffJ Eur Commun 1977; L105: 10. European Communities. Off J Eur Commun 1980; L229: 11.