Thyroid-Stimulating Antibodies Predict Hyperthyroidisma
Terry F. Davies, MD, FRCP, FACE
Mount Sinai School of Medicine
New York, New York
 |
Introduction
|
---|
OVER 50% of patients with hyperthyroid
Graves disease relapse after a 12-month course of antithyroid drugs,
the actual percentage varying among populations and with their iodine
intake. The measurement of TSH receptor antibodies (TSHR-Abs) by
competitive radioreceptor assay or by thyroid cell bioassay (1) in
patients with Graves disease can be a useful predictor of relapse and
remission (2). However, controversy has surrounded the clinical utility
of TSHR-Ab measurements ever since my colleagues and I published the
first report of their use in the Lancet in 1977 (2). In that study, we
analyzed the usefulness of TSHR-Abs to predict relapse and remission
after 6 months of antithyroid drug treatment and found that highly
positive titers of the autoantibody were indeed useful predictors of
recurrence (Fig. 1
). Similar data had
been published previously using indirect measurements of TSHR-Absfor
example, the pioneering work of Alexander, who had showed clearly that
the failure of T3 to suppress radioiodine uptake increased the chances
of recurrence because a non-TSH thyroid-stimulator of radioiodine
uptake was present (3). Similar studies continue to confirm such
observations (4) of the failure to suppress T4 levels with T3 therapy.
The assay for TSHR-Ab that we had used at that time, the original
simple thyroid membrane radioreceptor assay of Smith and Hall (5), was
new, but the conclusions were similar to the data of the time. The
presence of an abnormal thyroid stimulator indicated a higher chance of
recurrence than in patients in whom such a stimulator could not be
detected. While a number of the publications that followed our report
confirmed this observation (6, 7, 8, 9, 10, 11, 12, 13, 14), others were less enthusiastic
(15, 16, 17, 18, 19, 20, 21). The original observation certainly had a number of problems.
The most serious was that the follow-up time for detecting recurrence
was only 6 months. But I had a fellowship to move to at the NIH, and
the study had to end! Furthermore, sensitive TSH assays were not
available to help us in the management of the patients. Nevertheless, I
believe time has confirmed the original general conclusions. The
presence of significant titers of biologically active TSHR-Abs
accurately predicts recurrence in up to 90% of such patients.

View larger version (11K):
[in this window]
[in a new window]
|
Figure 1. The original data showing prediction of recurrent Graves disease after 6 months treatment with antithyroid drugs (combined carbimazole and thyroxine). Data were expressed as a TSHR-Ab Index, where a low index indicated high titers of TSHR-Abs. On the left are shown those patients who remained in remission, and on the right are those patients who had recurrence defined as increased T3 and T4 levels. (Data redrawn from ref. 2).
|
|
 |
Interpreting the medical literature
|
---|
As discussed above and elsewhere, there have been many
publications examining the clinical usefulness of TSHR-Ab measurement.
I would like to focus here on a recent meta-analysis published in
JCEM that has not, to my knowledge, been discussed at length
in the literature (22). The conclusion from this attempted
meta-analysis of 18 different reports was that "the available methods
for TSHR-Ab do not allow sufficiently high prediction of relapse or
remission after antithyroid drugs for the individual patient." I find
this conclusion inconsistent with the data presented by the authors and
other commentators (23).
The meta-analysis included only certain of the available studies
according to a number of clear criteria:
- TSHR-Ab measurements were available at the end of
antithyroid drug treatment.
- The follow up period after stopping therapy was at least 1
yr.
- Individual rather than grouped data were available.
- No thyroid hormone therapy was given after stopping antithyroid
drugs.
Each of these criteria was appropriate and strictly adhered to in
the studies selected. However, no criteria at all were applied to the
assay of TSHR-Abs except to divide the data into studies with
radioreceptor assays and studies with bioassays. The overall assumption
of the meta-analysis, therefore, was that all the different assays used
in the 18 different studies were of equal sensitivity, equal
precision and, most importantly, equal
specificity. However, many of the studies used their own
"in house" methods for the measurement of TSHR-Abs, sometimes for
the first time. We lack, therefore, essential data on the assay
performances, which would make a meta-analysis possible to interpret.
For example, in one study (24) only 3 out of 184 patients had
detectable TSHR-Abs at the end of antithyroid drug treatment. In
contrast, in another study 142 out of 391 patients had TSHR-Abs
detected (18). In addition, the actual titer of antibody was not taken
into consideration in any of the studies analyzed. Borderline positive
results were given the same weight as very high titers. Furthermore,
what did the different authors mean by remission? Was the TSH
normalized in these patients? Relapse and remission were also not
defined in this meta-analysis of relapse and remission.
The body of literature on this subject is, therefore, difficult
to interpret without careful scrutiny of each study, and
inclusion/exclusion criteria would need to be so strict that there
would be insufficient data to perform a meta-analysis. Yet we have to
conclude that some reports have clearly shown that more than 90% of
higher titer TSHR-Ab positive patients had a recurrence of Graves
disease after withdrawal of antithyroid drugs (13, 25), while this was
not seen in other reports (18). How can we explain these disparities? I
would like to begin with the only logical conclusion there can be. The
presence of TSH receptor antibodies with stimulating activity in the
serum of a patient with recent hyperthyroid Graves disease must
influence thyroid function, unless there is some interference in the
transduction of such a signal.
 |
Potential factors interfering with the action of TSHR-Abs
|
---|
There are a number of possible reasons for thyrotoxic recurrence
after antithyroid drugs not being as predictable as my logic
suggests:
1. Antibody quantity and affinity. In theory it is possible
that the level of TSHR-Abs could be so insignificant or their affinity
so low that their influence on the TSH receptor would be negligible.
However, under such circumstances it is unlikely that the TSHR-Abs
would be detectable using currently available assays, which remain
relatively insensitive to TSH equivalents of less than 10 mU/L.
Furthermore, even a slightly increased TSH level is known to be capable
of increasing thyroid hormone levels, as seen in patients with thyroid
hormone resistance. Hence, the very presence of detectable TSHR-Abs
with stimulating activity should lead to hyperthyroidism unless the
assay systems employed are providing false positive results or unless
there is a target defect (i.e. thyroid dysfunction).
2. Antibody quality. Not all TSHR-Abs are stimulating
antibodies. Blocking TSHR-Abs have been well characterized in many
laboratories, and some patients with Graves disease have been shown
to have both types of TSHR-Abs present in their serum (26). Of course,
in order for hyperthyroidism to develop there must be a lower
concentration or affinity of the blocking variety (27, 28, 29, 30). This has
led to discussions on the factors responsible for the balance between
stimulating and blocking antibodies and the differing methods for
measuring TSHR-Abs, discussions that have occupied clinical
thyroidologists for many years. The technique used for the measurement
of biologically active TSHR-Abs is indeed critical. Thyroid cell assays
have been replaced with more sensitive TSH receptor transfected Chinese
Hamster Ovary (CHO) cell stimulation assays (31) and remain the only
in vitro way to distinguish stimulating from blocking
activity. Many argue that it is never possible to infer the biological
activity of TSHR-Abs from competitive binding assays, and in some
circumstances this is certainly true. However, if the patient is
already thyrotoxic or has recently been thyrotoxic, it does not take a
great brain to deduce that the patients TSHR-Abs are predominantly of
the thyroid-stimulating variety. The real question, therefore, is
whether TSHR-Abs change from stimulating to blocking during the course
of antithyroid drug treatment. The literature to date is relatively
unhelpful in this regard, but this appears to be an unusual event (32, 33). Hence, the measurement of TSHR-Abs by an expensive bioassay is not
necessary for everyday clinical practice in patients with recent
hyperthyroid Graves disease.
3. Problems with thyroid function. There are two major
potential disturbances in the thyroid reserve that may prevent an
appropriate thyroid hormone response to TSHR-Abs of the stimulating
varietyiodine deficiency and autoimmune thyroiditis (34).
Iodine deficiency remains the major cause of global thyroid disease,
and moderate iodine deficiency continues to affect thyroid function
even in many parts of Western Europe such as Germany, where 15% of the
population may have iodine deficiency goiters (35). It is not
surprising, therefore, that low titers of TSHR-Abs would have
difficulty inducing hyperthyroidism in certain parts of the world
because of the lack of intrathyroidal iodine. However, in countries
such as the USA, Japan, and Korea, where there are high iodine diets,
this restriction should not apply. This certainly complicates the
meta-analysis from Feldt-Rasmussen (22), where many of the studies were
from countries with mild iodine deficiency.
Similarly, the onset of significant autoimmune thyroiditis would reduce
the thyroid reserve considerably and would require enhanced thyroid
stimulation to raise thyroid hormone levels above normal. Hence, the
presence of TSHR-Abs would not always predict the hyperthyroid state.
No data on the onset of hypothyroidism was provided for the 55 patients
in remission with detectable TSHR-Abs in large multicenter studies (18)
or in any of the other published prediction studies, but the geographic
location of the patients would suggest that iodine deficiency was more
important that thyroiditis. Furthermore, it appears to take many years
for thyroiditis to induce thyroid failure in some Graves patients
(36). We have all seen examples where this can happen much more
suddenly during treatment, but it is unusual.
4. Assay sensitivity. I believe the major problem today, with
easily available and relatively precise and specific assays, is not
false positive TSHR-Ab results but the large group of patients who
develop hyperthyroidism in the absence of detectable TSHR-Abs. Clearly,
in these patients a lack of thyroid reserve is not an issue. Accepting
the evidence that the presence of TSHR-Abs often predicts recurrence in
the presence of iodine sufficiency, does their absence predict
remission? Many studies of this problem have shown that a negative
result, as so often in medicine, is of little help to the clinician. Of
course the loss of previously detectable TSHR-Abs indicates that the
immune system may have become more tolerant, but such patients may
still have a high recurrence rate after discontinuing antithyroid
drugs. Is this due to problems with detecting low levels of TSHR-Ab?
Following the earlier discussion, there is much evidence to suggest
that only small increases in TSH (and, therefore, TSHR-Ab) lead to
thyroid hyperfunction. However, as also mentioned earlier, the
bioassays and radioreceptor assays used for the measurement of TSHR-Abs
are insensitive and/or imprecise to TSH concentrations of less than 10
mU/L. Low levels of receptor antibody may not be detected. In addition,
thyroid autoantibodies in the circulation are merely a reflection of
their intrathyroidal production. For example, patients with
Hashimotos disease and negative anti-Tg and anti-TPO still have
intrathyroidal B cells capable of secreting such antibodies (37). No
doubt the same applies to patients with Graves disease and negative
TSHR-Abs. Hence, assay sensitivity is likely to be the major factor in
the analysis of recurrent hyperthyroidism. Studies based on in-house
assays of dubious validity from inexperienced investigators are likely
to accentuate this intrinsic problem. Of course TSHR-Ab negative
patients may be patients who do not have Graves disease. The
diagnostic criteria used in published studies should also be closely
scrutinized.
 |
The cost effectiveness factor
|
---|
One of the most common arguments against the clinical use of
TSHR-Abs has been that the test is not cost-effective. In the United
States, laboratory costs are not reflected in the cost of the test
performed, but are based more on what the market will pay. In essence,
the radioreceptor assay for TSHR-Abs need only cost $1020, but in
practice the patient is more likely to pay $7080. Under laboratory
capitation contracts, this problem is eliminated. Furthermore, both
costs are low compared with testing methods in other diseases and
specialties. Assuming that all Graves patients are tested for
TSHR-Abs at diagnosis, then only those patients with higher titers
(>20% binding inhibition) need be tested again at the end of their
antithyroid drug regimen. Those patients who remain positive should
continue on antithyroid drugs for another 612 months before
retesting, as shown in the algorithm
(Fig. 2
). With a 90% prediction success
in positive patients the cost-effectiveness of this approach is self
evident. However, if only 5% of patients are TSHR-Abs positive at the
end of a course of antithyroid drug treatment, then this approach may
be impractical. This is not the clinical experience in most centers.
Physicians using this approach with antithyroid drug regimens of 18
months or more may find many of their patients to be TSHR-Ab negative
and the test unhelpful as many such patients will remit after such
long-term treatment. However, in my experience, a majority of patients
can be withdrawn from treatment much earlier, and those in whom such
discontinuation should be avoided can be identified by using the
TSHR-Ab testing algorithm (Fig. 2
). Nothing can be more cost effective
than treating patients for as short a time as possible while avoiding
unnecessary thyrotoxic recurrences.

View larger version (24K):
[in this window]
[in a new window]
|
Figure 2. A suggested clinical pathway for the treatment of Graves disease with antithyroid drugs using the TSHR-Ab test as an aid to decision making. In areas of iodine deficiency or where drugs are used for very prolonged periods, the pathway may need to be adjusted.
|
|
 |
The pregnancy conundrum
|
---|
The validity of predicting neonatal Graves disease using the
measurement of biologically active TSHR-Abs in the mother has been well
documented over many years (38, 39). The rarity of this problem
demonstrates the success of modern medical care with the early
treatment of Graves disease. This is, however, the one clinical
situation where the bioactivity of the TSHR-Abs may need to be known to
predict the neonatal thyroid function (40). Luckily, pregnancy is a
time when thyroid autoantibodies generally decrease in titer, due to
the secretion of trophoblast factors, which are immunosuppressive, and
bioactivity assessment is not necessary in many patients. However, in
those Graves patients with higher titers of TSHR-Ab by radioassay in
the third trimester, particularly those patients who have not been
recently hyperthyroid, a subsequent bioassay will allow the prediction
of neonatal hyperthyroidism or blocking antibody-induced hypothyroidism
(27, 41). This approach applies most importantly to women previously
treated for Graves disease and who are euthyroid or on thyroxine
replacement but have persisting titers of TSHR-Abs.
 |
Summary
|
---|
Many physicians would pay dearly to obtain a marker for the
disease that interests them most. Those of us interested in caring for
patients with Graves disease have such a marker. Knowing the titer of
TSHR-Abs in patients is useful in the prediction of thyrotoxic
recurrence after antithyroid drugs. A careful review of the literature
shows strong evidence for the observation that their accurately
measured presence predicts hyperthyroidism in more than 90% of cases
in iodine-sufficient areas. The assay of TSHR-Abs, therefore, remains a
most useful addition to the clinical armamentarium and a low-cost help
in treatment planning. Assays to distinguish blocking from stimulating
antibodies remain expensive but are only clinically important in
patients with pregnancy (40). The major hurdle remains one of
increasing the sensitivity of the available assays for TSHR-Ab so that
their usefulness can be applied successfully to an even greater
proportion of patients with Graves disease.
 |
Footnotes
|
---|
Address all correspondence regarding these controversies and requests
for reprints to: Terry F. Davies, M.D., Mount Sinai Medical Center, Box
1055, 1 Gustave L. Levy Place, New York, New York 10128.
a Supported in part by: the National Institutes
of Health Grants DK52464, DK35764, and DK45011 from NIDDK. Please note
that Dr. Davies is a Board member of the Kronus company, San Clemente,
California, which sells assays for the measurement of thyroid
autoantibodies including TSH receptor autoantibodies. The views
expressed are those of the author only.
Accepted May 20, 1998.
 |
References
|
---|
-
Davies TF. 1996 The pathogenesis of Graves
disease. In: Braverman LE, Utiger RD, eds. The thyroid: a fundamental
text. Philadelphia: Lipincott; 525536.
-
Davies TF, Yeo PP, Evered DC, Clark F, Smith BR, Hall
R. 1977 Value of thyroid-stimulating-antibody determinations in
predicting short-term thyrotoxic relapse in Graves disease. Lancet. 1:11811182.[Medline]
-
Alexander W, McLarty DG, Robertson J, et al. 1970 Prediction of the long-term results of antithyroid drug therapy for
thyrotoxicosis. J Clin Endocrinol. 30:540543.[Medline]
-
Prakash R. 1996 Prediction of remission in
Graves disease treated with long-term carbimazole therapy: Evaluation
of technetium-99m thyroid uptake and TSH concentrations as prognostic
indicators. Eur J Nucl Med. 23:118122.[Medline]
-
Smith BR, Hall R. 1974 Thyroid-stimulating
immunoglobulins in Graves disease. Lancet. ii:427429.
-
Wilson R, McKillop JH, Pearson DW, Cuthbert GF, Thomson
JA. 1985 Relapse of Graves disease after medical therapy:
predictive value of thyroidal technetium-99m uptake and serum
thyroid-stimulating hormone receptor antibody levels. J Nucl
Medicine. 26:10241028.[Abstract]
-
Zakarija M, McKenzie JM, Banovac K. 1980 Clinical
significance of assay of thyroid-stimulating antibody in Graves
disease. Ann Intern Med. 93:2832.[Medline]
-
Teng CS, Yeung RTT. 1980 Changes in
Thyroid-Stimulating Antibody Activity in Graves Disease Treated with
Antithyroid Drug and Its Relationship to Relapse: A Prospective Study. J Clin Endocrinol Metab. 50:144147.[Abstract]
-
Edan G, Massart C, Hody B, Poirier JY, Le Reun M,
Hespel JP, Leclech G, Simon M. 1989 Optimum duration of
antithyroid drug treatment determined by assay of thyroid stimulating
antibody in patients with Graves disease. Brit Med J. 298:359361.[Medline]
-
Schleusener H, Finke R, Kotulla P, Wenzel KW, Meinhold
H, Roedler HD. 1978 Determination of thyroid stimulating
immunoglobulins (TSI) during the course of Graves disease. A reliable
indicator for remission and persistence of this disease? J Endocrinol
Invest. 2:155160.
-
Karlsson FA, Dahlberg PA. 1980 Thyroid stimulating
antibodies (TSAb) in patients with Graves disease undergoing
antithyroid drug treatment: indicators of activity of disease. Clin
Endocrinol. 14:579585.
-
McGregor AM, Rees-Smith B, Hall R, Petersen MM, Miller
M, Dewar P. 180 Prediction of relapse in hyperthyroid Graves
disease. The Lancet. 11011103.
-
Wilson R, McKillop JH, Henderson N, Pearson DW, Thomson
JA. 1986 The ability of the serum thyrotropin receptor antibody
(TRAb) index and HLA status to predict long-term remission of
thyrotoxicosis following medical therapy for Graves disease. Clin
Endocrinol. 25:151156.[Medline]
-
Cho B, Shong MH, Yi KH, Lee HK, Koh C-S, Minard K. 1992 Evaluation of serum basal thyrotropin levels and thyrotrophin
receptor antibody activities as prognostic markers for discontinuation
of antithyroid drug treatment in patients with Graves disease. Clin
Endocrinol. 36:585590.[Medline]
-
Gossage AAR, Crawley JCW, Copping S, Hinge D, Himsworth
RL. 1982 Graves disease: thyroid function and immunologic
activity. J Nucl Med. 23:973977.[Abstract]
-
Croxson MS, Lim TMT, Graham FM, Ibbertson HK. 1980 Thyrotrophin displacement activity of serum immunoglobulins in health
disease. Aust N Z J Med. 10:151156.[Medline]
-
Schleusener H, Schwander J, Holl G, et al. 1987 Do
HLA-DR-typing and measurement of TSH-receptor antibodies help in the
prediction of the clinical course of Graves thyrotoxicosis after
antithyroid drug treatment? Acta Endocrinol (Copenh). 281:318324
[Suppl].
-
Schleusener H, Schwander J, Fischer C, et al. 1989 Prospective multicentre study on the prediction of relapse after
antithyroid drug treatment in patients with Graves disease. Acta
Endocrinol (Copenh). 120:689701.[Medline]
-
Young ET, Steel NR, Taylor JJ, et al. 1988 Prediction of remission after antithyroid drug treatment in Graves
disease. Q J Med. 250:175189.
-
deBruin TW, Bolk J, Bussemaker J, et al. 1988 Graves disease: immunological and immunogenetic indicators of
relapse. Br Med J. 296:12921295.[Medline]
-
Weetman A, Ratanachaiyavong S, Middleton GW, et al. 1985 Prediction of outcome in Graves disease after carbimazole
treatment. Quarterly Journal of Medicine 228:409419.
-
Feldt-Rasmussen U, Schleusner H, Carayon P. 1994 Meta-analysis evaluation of the impact of thyrotropin receptor
antibodies on long term remission after medical therapy of Graves
disease. J Clin Endocrinol Metab. 78:98102.[Abstract]
-
Braverman LE. 1995 Thyroid. In: Bagdade JD, ed.
Yearbook of endocrinology. St. Louis, MO: Mosby Yearbook; 266267.
-
Murakami M, Koizumi Y, Aizawa T. 1988 Studies of
thyroid function and immune parameters in patients with hyperthyroid
Graves disease in remission. Journal of Clinical Endocrinology and
Metabolism. 66:103108.[Abstract]
-
Rapoport B, Greenspan F, Filetti S, Pepitone M. 1984 Clinical Experience with a Human Thyroid Cell Bioassay for
Thyroid-Stimulating Immunoglobulin. J Clin Endocrinol Metab58
:332338.
-
Kraiem Z, Lahat N, Glaser B, Baron E, Sadeh O, Sheinfeld
M. 1987 Thyrotropin receptor blocking antibodies: incidence,
characterization, and in-vitro synthesis. Clin Endocrinol. 27:409421.[Medline]
-
Zakarija M, McKenzie JM, Eidson MS. 1990 Transient
neonatal hypothyroidism: characterization of maternal antibodies to the
thyrotropin receptor. J Clin Endocrinol Metab. 70:12391246.[Abstract]
-
Davies TF, Platzer M, Schwartz A, Friedman E. 1983 Functionality of thyroid-stimulating antibodies assessed by
cryopreserved human thyroid cell bioassay. J Clin Endocrinol
Metab. 57:10211027.[Abstract]
-
Creagh F, Teece M, Williams S, Didcote S, Perkins W,
Hashim F, Rees-Smith B. 1985 An analysis of thyrotrophin receptor
binding and thyroid stimulating activities in a series of Graves sera.
Clinical Endocrinology 23:395404.
-
Creagh F, Howells RD, Williams S, et al. 1986 IgG
TSH receptor antibody activity in Graves disease: a study of TSH
agonist and antagonist activities by isolectric focusing. Clinical
Endocrinology 24:7988.
-
Kim MR, Faiman C, Hoogwerf BJ, Gupta MK. 1997 Thyroid-stimulating antibody assay using a human thyrotropin receptor
transfected cell line: Relationship to clinical features of Graves
disease. Endocr Pract. 3:337343.
-
Kasagi K, Lida Y, Koshishi J. 1986. Paired
determination of thyroid-stimulating and TSH binding inhibitory
activities in patients with Graves disease during antithyroid drug
treatment. Acta Endocrinol (Copenh). 111:474480.
-
Tamai H, Hirota Y, Kasagi K. 1987 The mechanism of
spontaneous hypothyroidism in patients with Graves disease after
antithyroid drug treatment. J Clin Endocrinol Metab. 64:718722.[Abstract]
-
Rees Smith B, McLachlan SM, Furmaniak J. 1988 Autoantibodies to the thyrotropin receptor. Endocr Rev. 9:106121.[Abstract]
-
European Thyroid Association. 1985 Goiter and
iodine deficiency. Lancet. i:12891293.
-
Wood LC, Ingbar SH. 1979 Hypothyroidism as a late
sequela in patients with Graves disease treated with antithyroid
agents. J Clin Invest. 64:14291436.[Medline]
-
Baker JR, Saunders NB, Wartofsky L, Tseng Y-C, Burman
KD. 1988 Seronegative Hashimoto thyroiditis with thyroid
autoantibody production localized to the thyroid. Ann of Intern Med. 108:2630.[Medline]
-
Munro DS, Dirmikis SM, Humphries H, Smith T, Broadhead
GD. 1978 The role of thyroid-stimulating immunoglobulins of
Graves disease in neonatal thyrotoxicosis. Br J Obstet Gynaecol85
:837843.
-
Zakarija M, McKenzie JM. 1983 Pregnancy-associated
changes in thyroid-stimulating antibody of Graves disease and the
relationship to neonatal hyperthyroidism. J Clin Endocrinol Metab. 57:10361040.[Abstract]
-
Davies TF. 1988 Thyroid-stimulating hormone
receptor antibodiesthe calm after the storm after the calm? Mayo Clin
Proc. 63:736739.[Medline]
-
Matsoura N, Yamada Y, Nohara Y, et al. 1980 Familial neonatal transient hypothyroidism due to maternal TSH-binding
inhibitor immunoglobulins. N Engl J Med. 303:738742.[Medline]