Clinical Endocrinology (1981) 15, 395-401

HYPERTHYROIDISM IN GESTATIONAL
TROPHOBLASTIC NEOPLASIA

R . J . N O R M A N , R . W. GREEN-THOMP S ON , I. JIALAL, W. P. S O U T T E R ,
N . L. PILLAY A N D S . M . J OU B E R T

Departments of Chemical Pathology and Obstetrics and Gynaecology, University

of Natal, Durban, South Africa

(Received I April 1981: reirised 20 May 1981: accepted 3 June 1981)

SUMMARY

The thyroid status of twenty-seven African patients with gestational tropho-

blastic neoplasia (GTN) was studied. Fifteen patients were found to be
biochemically hyperthyroid (eight patients with choriocarcinoma; seven with
hydatidiform mole). Of these fifteen patients, nine were clinically thyrotoxic.
The most serious complication of thyrotoxicosis was life-threatening acute
pulmonary oedema with associated cardiac failure. It was found that when
serum levels of the human chorionic gonadotrophin (hCG) reached a level of
about 0.1 x lo6 iu/l, thirteen of sixteen patients were biochemically hyper-
thyroid; at serum levels of 0.3 x lo6 iu/l of hCG most patients were clinically
thyrotoxic. A feature of hyperthyroidism associated with GTN is that whereas
T4 is invariably raised the T3 :T4 ratio tends to be low (0.015 f005); rT3 :T3
ratios were high in this group. TSH levels were not increased.

Since the first recorded observations of an association between hyperthyroidism and

gestational trophoblastic neoplastia (GTN), several descriptions of altered thyroid
function in this disease spectrum have been published (Dowling et al., 1960; Galton et al.,
1971; Nagataki et al., 1977). These have emphasized the role of a placental thyrotrophic
factor, isolated from sera and tumours, which stimulates the thyroid gland and is distinct
from both pituitary thyrotrophin (TSH) and thyroid stimulating immunoglobulins
(Higgins & Hershman, 1978). Although overt hyperthyroidism with classical features of
thyrotoxicosis has been described in GTN, reports from several centres (Ode11et al., 1963;
Nagataki et al., 1977) have emphasized that most patients with hydatidiform mole and
choriocarcinoma have few, if any, clinical signs of excess thyroid secretion despite
elevated blood levels of thyroxine and triiodothyronine. In addition, the percentage of
patients with biochemical hyperthyroidism has varied considerably in several series.
As trophoblastic disease is relatively common in many developing communities, we

Correspondence: Dr R. J. Norman, Department ofchemical Pathology, University of Natal, PO Box 17039.

Congella 4013, Durban, South Africa.
0300-0664/81/1000-0395$02.00 0 1981 Blackwell Scientific Publications
395
396 R.J. Norman et al.
report on the prevalence, clinical features and thyroid status of patients with this disease
at King Edward VIII Hospital, Durban, South Africa.

PATIENTS A N D M E T H O D S
Twenty-seven consecutive black patients with trophoblastic disease admitted to the
gynaecological wards of King Edward VIII Hospital over a 12-month period were
studied. Fifteen had choriocarcinoma and twelve hydatidiform mole as determined by
clinical assessment, ultrasonography, currettage and histopathological study of surgical
specimens. All patients were examined on admission to exclude hyperthyroidism and a
chest x-ray and electrocardiogram (ECG) were performed. Clinical hyperthyroidism was
considered to be present when some or all of the following clinical features were present:
tachycardia, excessive sweating, fine tremor, heat intolerance, proximal myopathy,
weight loss, hyperactive reflexes.
Blood was taken from patients on admission and on every second day thereafter for
biochemical assessment as described below. Potassium iodide drops orally with a
beta-adrenergic blocker (propanolol 120 mg/day or sotalol hydrochloride 160 mg/day)
were used where clinical and laboratory evaluations suggested thyrotoxicosis. Evacuation
of hydatidiform moles was by gentle currettage under general anaesthesia while treatment
of choriocarcinoma was by cytotoxic therapy with one or a combination of the following
drugs: methotrexate, hydroxyurea, actinomycin D, vincristine, 6-mercapto-purine,
cyclophosphamide. Folinic acid rescue was instituted 40 h after methotrexate infusion.

Biochemical investigations
Total thyroxine (T4), 3,5,3’ triiodothyronine (T3), human chorionic gonadotrophin
(hCG), 3,3’,5’ ‘reverse’ triiodothyronine (rT3) and thyroid stimulating hormone (TSH)
were measured by radioimmunoassay using reagents supplied by the Radiochemical
Centre, Amersham, Bucks, UK, (T4, T3), Biodata Serono, UK, (hCG & rT3) and
Diagnostic Products Corporation, Los Angeles, California (TSH). hCG was determined
using an antibody raised against the beta subunit of hCG and results expressed in units of
the 2nd International Standard for hCG. Thyroid hormone binding capacity was
measured by the Thyopac-3 (Amersham) and a free thyroxine index calculated. Cross
reaction of hCG with TSH was 0.1 x 106-(0.1 x Within and between assay
coefficients of variation for all assays were less than 10%. Biochemical hyperthyroidism
was adjudged present if two or more of the following features were present: serum T4
greater than 180.0nmol/l; nofmal or decreased thyroid hormone binding capacity; serum
T3 greater than 3.5 nmol/l; free thyroxine index greater than 190 nmoI/l (Burrow, 1978).

RESULTS
Clinical features
Fifteen of the twenty-seven patients admitted to the study were hyperthyroid as judged
by the stated criteria. Eight of these had choriocarcinoma and seven hydatidiform mole.
The average age of those with choriocarcinoma was 30 and those with hydatidiform mole
29 years. The most common presenting features were vaginal bleeding (24/27) and
abdominal pain (20/27).
Clinical manifestations of increased thyroid hormone secretion were seen in nine
 Hyperthyroidism in gestational trophoblastic neoplasia 397
Table 1. hCG and thyroid hormone concentrations in patients with gestational tropho-
blastic neoplasia-choriocarcinoma

hCG T4 FT1 T3 rT3 T3:T4 rT3:T3

Choriocarcinoma (iu/ I x lo6) (nmolil) (nmolil) (nmolil) (nmolil) ratio ratio

Biochemical
hyperthyroidism
1* 9.00 424.7 499.4 7.7 2-31 0.0I8 0-30
2*t 4.00 311.5 332.0 6.1 2.39 0.010 0.39
3*t 1.37 772.2 1193.0 8.9 4.60 0.012 0.52
4* 0.75 272.8 298.5
5* 0.48 23 1.7 261.3 4.3 0.019
6t 0.38 218.8 225.2 1.9 0.009
7t 0-17 280.6 293.4 1.1 1.90 0.004 1.72
8 0.10 289.6 272.8 6.3 1.42 0.022 0.25
Euthyroid
choriocarcinoma 0.111 129.8 110.1 1.57 0.62 0.013 0.53
(n= 7) range (0.01- (9443 (66.9- (0.5- (0.30- (0.055- (0.2cL
0.19) 171.2) 144.1) 2.2) 1.26) 0.19) 1.36)

* Clinically hyperthyroid.
t Died.

patients and included tachycardia, tremor, sweating and proximal myopathy. Cardiac
arrhythmias, other than supraventricular tachycardia, were absent and the characteristic
eye signs of Grave's disease were not seen. Goitre, other than that expected in pregnancy,
was not in evidence.
The most serious clinical presentation in hyperthyroidism due to GTN was a life
threatening acute pulmonary oedema and respiratory distress accompanied by other
features of cardiac failure. This was frequently refractory to standard medical manage-
ment despite the use of high dosage diuretics and, in one patient, venesection.
Beta-adrenergic blocking drugs were avoided in these situations until the acute phase had
subsided as a result of reduction of tumour mass by evacuation or chemotherapy and the
reduction of thyroid secretion by potassium iodide administration.
Five of the fifteen patients died during the course of treatment. Two patients with
choriocarcinomas, which were resistant to chemotherapy (Nos 6 & 7) died of widespread
metastases; both these patients became biochemically hyperthyroid after admission. Two
other patients with choriocarcinoma (Nos 2 & 3) died shortly after admission. Their
thyroid function tests were consistent with hyperthyroidism; permission for post-mortem
could not be obtained. One patient with hydatidiform mole, who was hyperthyroid on
admission, was allowed home when biochemically euthyroid. She collapsed and died of
an intracerebral haemorrhage while returning to hospital.

Biochemical evaluation
Results of thyroid function tests are shown in Tables 1 and 2. Serum thyroxine
concentrations were invariably elevated in the biochemically hyperthyroid group and
there was a strong correlation between levels of hCG and thyroxine values above 180
nmol/l (r = 0.64, P < 0.0001). Serum T3 was also raised in most patients but was low in the
398 R . J. Norman el al.
Table 2. hCG and thyroid hormone concentrations in patients with gestational tropho-
blastic neoplasia-hydatidform mole

hCG T4 FTI T3 rT3 T3:T4 rT3:T4

Hydatidiform mole (iu/l x lo6) (nrnolil) (nrnolil) (nmol/l) (nmol/l) ratio ratio

Biochemical
hyperthyroidism
9*f 1.80 537.9 456.9 8.5 1.85 0.016 0.22
lo* 0.78 242.0 202.5 5.5 1.85 0.023 0.34
11 0.58 22 1.4 193.1 3.8 2.23 0.017 0.59
12* 0.5 1 386.1 350.1 5.5 0.014
13 0.14 170.3 217.5 3.6 1.51 0,013 0.42
14* 0.096 267-7 293-4 3.6 0.014
15 0.065 292.1 293.4 2.9 2.08 0~010 0.71
Euthyroid
mole 0.016 143.6 114.8 2.3 0.32 0,018 0.15
( n= 5 ) range (0.001- (92.7- (74.6 (1.8- (0.16 (0.015- (0.07-
0.033) 136.4) 136.4) 2.5) 0.52) 0,020) 0.22)
Upper normal pregnant
reference range 180.0 193.0 3.5 0.60

* Clinically hyperthyroid
t Died.

terminal stages of the disease, particularly in two patients with biochemical hyperthy-
roidism where it was measured. TSH concentrations were all low (less than 4 mu/l) and
did not show significant changes during therapy.
There was a close association between absolute levels of immunoreactive hCG and the
presence of hyperthyroidism (Table 3) in that thyrotoxicosis was present in all patients
when hCG values were above 0.3 x lo6 iu/L Thirteen of sixteen patients with concentra-
tions greater than 0.1 x lo6 iu/l were biochemically hyperthyroid.
Three main patterns of thyroid response to therapy were observed.
1 A rapid fall in T4, T3 and rT3 levels, a rise in the thyroid hormone binding capacity
accompanied by a decrease in hCG (three choriocarcinomas and all hydatidiform moles).
2 Where the tumour was partially resistant to the initial cytotoxicdrugs used, theT4 and

Table 3. Serum hCG concentrations in relation to biochemical

and clinical hyperthyroidism

hCG Number in Biochemical Clinical features

(iui1 x lob) group hyperthyroidism of hyperthyroidism

< 0.05 7 0
04-0.099 4 2
0.10-0-299 6 3
0.30-0.60 4 4
> 0.60 6 6
27 15
 Hyperthyroidism in gestational trophoblast ic neoplasia 399
T3 concentrations remained at plateau levels for several days or weeks until alternative
effective therapy was instituted.
3 In patients admitted with choriocarcinoma resistant to chemotherapy there was rapid
progress from the euthyroid state to biochemical hyperthyroidism with a rise in hCG
blood levels followed by clinical thyrotoxicosis and death (two choriocarcinomas).
A notable feature in the two patients with the response noted under (3)was a drop in T3
values in the terminal stages and, in one case, where rT3 was measured as well, a marked
increase in rT3 values. The drop in T3 level and the rise in rT3 level was associated with
clinical improvement of the thyrotoxic state.
Patients with molar pregnancy and thyrotoxicosis returned to a clinical euthyroid state
about 5 days after evacuation of the mole and in two patients with choriocarcinoma
resistant to cytotoxic drugs, the thyroid status changed from euthyroid to biochemical
hyperthyroidism followed by clinical hyperthyroidism about 5 days later.

DISCUSSION
It is obvious from the number of patients with GTN who presented at the gynaecological
wards over a period of 12 months that choriocarcinoma and hydatidiform mole are
relatively common amongst the local population. This accords with the fact that GTN is
more frequent in many developing countries and in certain racial groups (Bagshawe,
1969; van Dongen, 1972).
The present study has confirmed that biochemical evidence of thyroid over-activity is
common in GTN whereas clinical signs and symptoms are seen less often (Ode11 et al.,
1963; Galton et al., 1971; Nagataki et al., 1977). At least fifteen of the twenty-seven
patients had definite biochemical evidence of hyperthyroidism whereas only nine had
clinical features of thyrotoxicosis. However, as is clear from experience with patients who
presented with tumours resistant to cytotoxic therapy, the natural history of the disease is
an initial euthyroid state, progressing to biochemical hyperthyroidism and ultimately to
clinical manifestations of thyrotoxicosis. This is further supported by data in Table 3.
Inasmuch as circulating hCG is a function of tumour size (Bagshawe, 1976)it is clear that
as blood levels of hCG increase beyond 0.05 x lo6 iu/l, biochemical hyperthyroidism
becomes more common and, when concentrations exceed 0.3 x lo6 iu/l, hyperthyroidism
is always present. It would thus appear that, in any population of patients with GTN, the
prevalence of either biochemical or clinical thyroid overactivity is a function of the stage
in the natural history of GTN and of the tumour mass. As many of the patients reported in
this series sought medical attention only when the disease was in its advanced stages, the
results should be seen in this light.
There is some interest relating to the observation that six patients with biochemical
evidence of hyperthyroidism exhibited no obvious clinical signs of the disorder.
Experience in the present study seems to suggest that changes in the level of thyroid
hormones in the blood will take at least 5 days to find clinically observable expression. It
may therefore be speculated that, in these six patients with biochemical hyperthyroidism,
sufficient time had not elapsed for clinical expression of the hyperthyroid state. An
alternative explanation may be the low T3:T4 ratios in GTN noted by Nagataki et al.
(1977). In the current study the ratio was 0.015+0.005 (mean* 1 SD) in GTN with
biochemical hyperthyroidism; 0.01 5 & 0.006 in GTN with euthyroid status; 0.025 & 0,006
400 R.J . Normun et al.
in Grave’s disease as calculated from fifteen cases seen in the diagnostic endocrine unit of
this hospital. Because T3 is the more active hormone, the significantly lower ratio of T3 to
T3 may, in part, account for the disproportion between clinical and biochemical
hyperthyroidism and may also explain the relatively mild clinical features seen in many
patients with markedly increased T4 levels. Yet a further consideration is the higher ratio
of rT3 to T3 in this group of patients. A high rT3:T3 ratio was particularly marked in
patient no. 7 who, in the terminal stages of her disease, showed remission of clinical signs
of hyperthyroidism associated with very high rT3 and low T3 levels. There were two
patients who had hCG blood levels of 0.29 x lohiu/l and 0.9 x lo6$1, respectively (Table
3) and may have been expected to exhibit abnormalities of thyroid function. Yet, they
were euthyroid biochemically and clinically. Notable, however, was the high rT3 levels in
these two patients, suggesting increased peripheral conversion of T4 to rT3 instead of T3.
Is laboratory evidence for hyperthyroidism in the absence of clinical features of any
significance? Local clinical experience indicates that the answer is yes. As the disease
progresses, cardiopulmonary complications may be life-threatening and may lead to
death. This is particularly true in the management of choriocarcinoma where levels of
hCG cannot be decreased as rapidly as in molar pregnancy. Cytotoxic resistance may lead
to a euthyroid patient with initial low blood levels of hCG becoming hyperthyroid rapidly
as the disease progresses. As hCG and bioassable thyrotrophin are highly correlated
(Nagataki et al., 1977) the level of hCG is a reliable marker of changes in thyroid
stimulation and should be measured regularly together with T4 and T3 in all patients
undergoing cytotoxic drug therapy. Routine prophylaxis and treatment of hyperthy-
roidism should include the use of oral or p a r e n t e d iodine therapy together with a
beta-blocking agent. Propanolol is the agent of choice because of its effects on the
peripheral conversion of T4 to T3 and rT3 (Feely et al., 1979).Other beta-blockers are not
as effective in this respect (Jones et al., 1980).
There has been extensive investigation into the identity of the trophoblastic thyrotro-
phin and recent evidence has emphasized the role of hCG itself as the thyroid stimulator
(Cave & Dunn, 1976). hCG is indistinguishable from the thyrotrophin extracted from
trophoblastic tumours and sera as judged by several physico-chemical criteria (Kecimer
er al., 1975); it binds to the mouse TSH receptor and generates cyclic AMP in animal
thyroid slices (Higgins & Hershman, 1978). In short, hCG fulfils many criteria for the
thyrotrophin of GTN due to its common alpha-chain structure with TSH and the
similarity of its beta-chain. Eiowever, despite these characteristics, Amir et al. (1980)
continue to caution against accepting hCG as the sole thyroid stimulator in tropho-
blastic disease. These authors were unable to demonstrate significant adenyl cyclase
activation or competitive binding for the TSH receptor by pure hCG in the human
thyroid membrane.
Irrespective of any controversy regarding the true nature of the thyrotrophic stimulus
in GTN, two facts are clear from the present study: hCG levels are closely correlated with
hyperthyroidism above indicated critical levels and are accordingly a predictor of thyroid
gland hyperactivity; thyroid stimulation is clearly not mediated via pituitary TSH as
manifested by the low or normal TSH levels seen in all patients. Whereas the TSH assay
employed is insensitive in the range of TSH concentrations reported in this study, it is
nevertheless adequate to demonstrate TSH-induced hyperthyroidism. Thus it should be
concluded that, on balance, the evidence presented here argues strongly for hCG as the
active thyroid stimulator.
 Hyper thyroidism in gestational trophoblast ic neoplasia 40 1
ACKNOWLEDGEMENTS

This work has been supported through the Preclinical Diagnostic Research Unit of the
South African Medical Research Council and a grant from the Medical School Research
Fund to R. W. Green-Thompson.

REFERENCES
AMIR,S.M., SULLIVAN, R.C. & INGBAR, S.H. (1980) In uitro responses to crude and purified hCG in human
thyroid membranes. Journal of Clinical Endocrinology and Metabolism, 51, 51-58.
BAGSHAWE, K.D. (1969) Choriocarcinoma. Edward Arnold, London.
BAGSHAWE, K.D. (1976) Immunological methods in the diagnosis and monitoring of tumours. In Medical
Oncology-Medical aspects of malignant disease (ed. K. Bagshawe), pp. 245-267. Blackwell Scientific
Publications, Oxford.
BURROW,G.N. (1978) Maternal-fetal considerations in hyperthyroidism. Clinics in Endocrinology and
Metabolism, 7, 115-125.
CAVE,W.T. & DUNN,J.T. (1976) Choriocarcinoma with hyperthyroidism: probable identity of the thyrotropin
with human chorionic gonadotrophin. Annals of Internal Medicine, 85, 6&63.
DOWLING,J.T., INGBAR,S.H. & FREINKEL, N. (1960) Iodine metabolism in hydatidiform mole and
choriocarcinoma. Journal of Clinical Endocrinology and Metabolism, 20, 1-1 I .
FEELY,J., ISLES,T.E., RATCLIFFE,W.A. & CROOKS,J. (1979) Propanolol, triiodothyronine, reverse
triiodothyronine and thyroid disease. Clinical Endocrinology, 10, 531-538.
GALTON, V.A., INGBAR, S.H., JIMENEZ-FONSECA, J. & HERSHMAN, J.M. (1971) Alterations in thyroid hormone
economy in patients with hydatidiform mole. Journal of Clinical Investigation, 50, 1345-1 354.
HERSHMAN, J.M. (1978) Use of thyrotrophin-releasing hormone in clinical medicine. Medical Clinics of North
America, 62, 313-324.
HIGGINS, H.P. & HERSHMAN, J.M. (1978) The hyperthyroidism due to trophoblastic hormone. Clinics in
Endocrinology and Metabolism, 7, 167-174.
JONES,M.K., JOHN,R. & JONES,G.R. (1980) The effect of oxyprenolol, acebutolol and propanolol on thyroid
hormones in hyperthyroid subjects. Clinical Endocrinology, 13, 343-348.
KENIMER, J.G., HERSHMAN, J.M. & HIGGINS,H.P. (1975) The thyrotropin in hydatidiform moles in human
chorionic gonadotrophin. Journal of Clinical Endocrinology and Metabolism, 40,482489.
NAGATAKI, S., MIZUNO,M., SAKAMOTO, S., IRIE,M., SHIZUME,K., NAKAO,K., GALTON, V.A., ARKY,R.A. &
INGBAR,S.H. (1977) Thyroid function in molar pregnancy. Journal of Clinical Endocrinology and
Metabolism, 44,254263.
ODELL, W.D., BATES, R.W., RIVLIN, R.S., LIPSETT,M.B. &HERTZ,R. (1963)Increased thyroid function without
clinical hyperthyroidism in patients with choriocarcinoma. Journal of Clinical Endocrinology and
Metabolism, 23,658-664.
VAN DONGEN, L. (1972) Hydatidiform mole and choriocarcinoma. In Gynaecology in Southern Africa (ed. G. P.
Charlewood), pp. 96122. Witwatersrand University Press, Johannesburg.