Clinical Endocrinology (1979) 10,443-449.

SERUM T 3 AND T 4 LEVELS I N PATIENTS WITH ANOREXIA NERVOSA SHOWING TRANSIENT HYPERTHYROIDISM

DURING WEIGHT GAIN

Department of Medicine, Addenbrooke’s Hospital, Hills Road, ambridge CB2 2QQ

(Received 16 August 1978; revised 5 October 1978;accepted 9 October 197%)

SUMMARY

Serum thyroxine (T4) and triiodothyronine (T3) levels were measured in a group of thirty-three patients with anorexia nervosa (AN) and compared with twenty-five control women presenting with hirsutism and twenty-one patients with primary myxoedema. T3 levels in the AN patients were significantly lower thanin the control subjects and in the patients with myxcedema while T4 levels were significantly higher than in the patients with hypothyroidism but significantly lower than in the control group. Seventeen anorexia patients had further T3 and T4 levels mea- sured following an arbitrary 25% weight gain and both levels had increased signifi- cantly. For individual patients, the absolute rise in T3 levels was significantly correlated with the rate of weight gain. Thirteen patients had serial T3 and T4 levels measured during their periods of weight gain. Eight of these subjects showed a gradual rise in T3 levels from subnormal or low normal levels to values in the upper normal range. Four subjects showed a distinct and self limiting overshoot of T3 levels and, associated with this, the patients had clinical features of mild hyperthyroidism.

The precise thyroid status of patients with anorexia nervosa has, until recently, been the sub- ject of considerable debate. Studies by Bliss& Migeon (1957), Crisp (1965), Rowland (1970) and Mills et ul. (1973) failed to demonstrate any consistent abnormality in thyroid function tests available at the time, except for a low basal metabolic rate (BMR).

Following the development of radio-immunoassay in the measurement of T4, T3 and thyrotrophin (TSH), several reports have appeared all confirming the presence of consistently low T3 levels and to a variable extent, T4 (Miyai etal. , 1975; Moshang etal., 1975; Croxson & Ibbertson, 1977; Boyer, 1978). Lundberg e t ul. (1970) reported a normal or delayed TSH response to thyrotrophin releasing hormone fTRH). Burman et ul. (1977) additionally reported raised reverse T3 (rT3) levels at the time of diagnosis. Only Croxson & Ibbertson’s paper showed a rise in T3 following weight gain based on a single repeat measurement. None of these papers has reported serial measurements in response to appropriate weight

Correspondence: Dr R. Moore, Anglesea Road Hospital, Ipswich, Suffolk.

0300-0164/79/0500-0443$02.00 o 1979 Blackwell Scientific Publications

443

444 R. Moore and I. H. Mills

gain which is the purpose of the present paper. The prospective study was prompted by the observation of a temporary clinically hyperthyroid state in the occasional patient who had gained weight rapidly in that she had a tachycardia, a fine tremor, was peripherally vaso- dilated and heat intolerant and was putting on a disappointingly small amount of weight in spite of a large dietary intake. Previous studies in a similar group of patients had shown an elevated BMR at this stage, but a normal free thyroxine index (Mills et al., 1973).

MATERIALS AND METHODS Patients

Between 1976 and 1978, thirty-three patients with anorexia nervosa referred to a general medical clinic run by one of us (IHM) with a special interest in this condition, were studied. All patients were female with a mean age of 19.9 years (range 12-36 years) and gave a typical history of food aversion and secondary amenorrhoea with a weight loss of at least 25% of original body weight (Feighner et al., 1972). All patients gave informed consent to the blood tests being performed and all had blood taken for T3 and T4 levels on presenta- tion. In eighteen patients, the condition was thought to be serious enough to warrant in- patient treatment and in this group of patients, further T3 and T4 estimations were per- formed in response to weight gain, In fourteen patients blood was taken at weekly intervals, but in the other four, insufficient data was collected serially although further blood was taken after a 25% weight gain. The patients were treated in a general medical ward with a reward and punishment regime supported by psychotherapy. All were given a tricyclic anti- depressant, usually amitriptyline in doses ranging from 50-300 mg/day in divided doses. A dietary intake of 3000-4000 Kcal was aimed for, containing approximately 55% carbo- hydrate and 15-20% fat.

The control group of patients matched for age and sex were twenty-five consecutive patients referred to an endocrine clinic with hirsutism. All were clinically euthyroid when first seen and all had thyroid hormone levels within the normal range for our laboratory (T3 1.0-2.8 nmol/l, T4 55-144 nmol/l). The initial T3 and T4 levels of twenty-one patients referred with primary myxoedema (TSH level greater than 10 mU/l) were also collected over this 2-year period.

Methods Serum concentrations of T3, T4 (Challand et al., 1975) and TSH (Hall et al., 1971)

were determined by radio-immunoassay. The TSH response to TRH was determined by the measurement of TSH at 0, 20, and 60 min after 200 pg of intravenous synthetic TRH (Roche, Welwyn Garden City).

Student’s paired 1 test was used for the statistical analyses.

RESULTS Serum T3

The initial serum T3 level (mean f SEM) was 1.00 +- 0.047 nmol/l in the 33 anorexia patients, significantly lower (PgO.001) than in the twenty-five control subjects, 1.95 f 0.08 nmol/l and the twentyane patients with primary hypothyroidism, 1.20 k 0.05 nmol/l (P<O.Ol).

T3 and T4 in anorexia nervosa 445

2.6 -

2.6 -

2.4 -

2.2 -

2.0 -

1.8 -

1.6- . ..-.

0

$ rc) 1.4 - - +

1.2 - .O. 0 .. 8

Serum 7'4 The initial serum T4 level in the thirty-three anorexia patients was 77.5 f 4.0 nmol/l,

significantly lower (P < 0.001) than the control group, 97.0 _+ 3.2 nmol/l but significantly higher (P< 0.001) than the patients with hypothyroidism, 26.3 f 3.5 nmol/l.

The clustering of the relative T3 and T4 levels in the three groups is shown in Fig. 1, with all the control values falling within the box representing the normal range for these values. The patients with primary hypothyroidism tended to have low T4 values with relatively well maintained T3 levels, whereas the converse was true in the anorexia patients.

In the anorexia patients there was no correlation between initial weight (mean f SEM) 38.5 f 1.04 kg, range 23-46 kg, and T3 or T4 levels on presentation.

The eighteen anorexia patients admitted for treatment had further thyroid hormone levels measured after an arbitrary 25% weight gain. T3 levels (2.4 f 0.12 nmol/l) in this group had risen significantly (P < 0.001) as compared to initial values, 1 .O * 0.04 nmol/l. T4 levels also rose, but to a lesser extent, 82.0 f 3.5 nmol/l compared with initial values of 71.5 * 3.9 nmol/l (P < 0.05). However, when examined individually, it is apparent from Fig. 2 that the extent of the T3 rise was dependent on the rate rather than the amount of weight gain since all these patients put on proportionately the same amount in relation to their initial weight. No such relationship was seen with the individual T4 values.

* * * *

* J( . * * *

* * * * * * *

* * * * * x o * 0

00 0 000 O* 0

0 * o 0" 0 n A " '.O-

0.8

0.6

0.4

0.2-

0 0 0 0

0 0 0 0

T . - 0 0 0

- 0 0

- 0

I I I I I I I 20 40 60 80 I00 I20 140

52.5-

45 - 0 x - - 2z 0 .- s

37.5

30

4 .(

-

-

-

3.1

2.(

c

\

0 - E

4 0.8

- I.(

0.8

0.’

0.1

O.!

0.8

0.:

-

-

-

0

4 0

30

r =-0.92

I I I I I I I 0 5 10 15 20 25 30

Weeks

Fig. 2. Correlation between absolute rise in T3 level (log arithmic scale) from time of diagnosis to period of arbitrary 25% weight gain and time taken to achieve this weight gain.

i3

2

\ -

E

m I- I: I

1 I I 1 I I I I I I I I

Weeks

2 4 6 8 10 12 14 16 18 20

Fig. 3. Gradual rise in T3 levels associated with weight gain in a 23yeardd girl, without significant change in T4 levels.

T3 and T4 in anorexia nervosa

7 .

6 -

5 . \ -

4 i . m c

3

L I I I I I I I I I 1

Weeks

I 2 3 4 5 6 7 8 9 10

140

120

100

2

1 ao

F b60

447

3

2 -

I -

40

20

Fig. 4. Exaggerated T3 rise in an 18yeardd girl in the overshoot group. Note associated levelling off of weight gain at this time.

3 2 . 5 ~ /.----.

T- I I I I I I I I I 0 I 2 3 4 5 6 7 8 9

Weeks

Fig. 5 . 12-yeardd girl in ‘overshoot’ gtoup.

448 R. Moore and I. H, Mills

Of the fourteen patients from those admitted who had weekly T3 and T4 measurement during theii period of weight gain, ten showed a gradual rise of the T3 levels from sub- normal or low normal values to levels in the high normal range. The T4 measurements re- mained low normal or rose only slightly. The detailed findings of a typical patient in this group is shown in Fig. 3. In contrast to these patients, four subjects showed a distinct and sustained overshoot of the T3 levels, falling back spontaneously towards high normal levels 2-4 weeks later. Two of these patients are illustrated in Figs. 4 & 5 and the findings were similar in the other two, reaching peak T3 values of 5.7 and 4.0 nmol/l. All of these patients were putting on weight rapidly in the few weeks before this overshoot and at the time of the high T3 levels, were clinically mildly hyperthyroid as described earlier. The rate of weight gain during this period diminished noticeably in spite of a continuing satisfactory dietary intake.

DISCUSSION

The findings of subnormal total T3 levels and low normal total T4 levels in anorexia nervosa at the time of hospital presentation have confirmed the previously recorded findings of other investigators. Croxson & Ibbertson (1977) reported a rise in the T3 levels from a mean of 1.15 to a mean of 1.42 nmol/l (P < 0.05) in six patients following a weight gain from 42.8 to 49.7 kg, a mean, but not an individual weight gain of 16% but they did not state the time taken to achieve this weight gain. The present findings, however, would suggest that this time is crucial in the interpretation of the T3 data since our results strongly suggest that the degree of T3 rise is directly proportional to the rate rather than the extent of weight gain and we have additionally shown that in four patients out of the fourteen studied serially, T3 levels clearly rose well above the normal range and these patients had clinical symptoms and signs appropriate to this temporary and self limiting hyperthyroid state. Probably the most important of these clinical features for not only these girls but also for the medical and nursing staff looking after them, is the temporary slowing down or even arresting of the characteristically excellent weight gain which they were achieving in the weeks prior to the T3 overshoot. Unless the cause is appreciated and explained to the patient and all those concerned with her care, there is a danger of demoralisation and false accusations to the patient with regard to occult vomiting or hiding of food.

Although we have shown a correlation between T3 rise and rate of weight gain it is likely that the direct correlation is in fact with the amount of food intake, particularly carbohydrates. In a study reported in abstract form by Darnforth et al. (1975), seven lean volunteers who were overfed with carbohydrate increased their T3 levels from 2.2 to 2.7 nmol/l (mean values, P < 0.02) with a corresponding increase in BMR, whereas four subjects who were overfed with added fat showed no change in the T3 levels or BMR. The standard refeeding treatment regime for anorexia patients involves a high carbohydrate intake but it is unlikely that a high fat diet as an alternative would be acceptable on the grounds of unpalatability.

It is unlikely that the tricyclic antidepressant drugs these patients were receiving during the course of the study had any direct effect on the results, since we have performed similar serial measurements on a group of depressed patients receiving similar dosages of these drugs and have observed no change in the T3 or T4 levels over a 12-week period (unpublished observations).

T3 and T4 in anorexia nervosa 449

ACKNOWLEDGEMENTS

We should like to express our thanks to the Department of Clinical Biochemistry, Adden- brooke’s Hopital, and in particular to Dr Gordon Challand, the head of the radioimmuno- assay department until January 1978, for performing the T3, T4 and TSH estimations.

This paper was presented to the meeting of the Endocrine Section of the Royal Society of Medicine which was held in Cambridge on 24 May 1978.

REFERENCES

BLISS, E.L. & MIGEON, C.J. (1957) Endocrinology of anorexia nervosa. Journal of Clinical Endocrino- logy and Metabolism, 17,766-774.

BOYER, R.M. (1978) Endocrine changes in anorexia nervosa. In The Medical Clinics ofNorth America, 62:2, pp. 297-304. W. B. Saunders, Philadelphia.

BURMAN, K.D., VIGERSKY, R.A., LORIAUX, D.L., STRUM, D., DJUH, Y.Y., WRIGHT, F.D. & WARTOFSKY, L. (1977) investigations concerning thyroxine deiodinative pathways in patients with anorexia nervosa. In Anorexia Nervosa, (ed. R. A. Vigersky), pp, 255-261. Raven Press, New York.

CHALLAND, G.S., RATCLIFFE, W.A. & RATCLIFFE, J.F. (1975) Semi-automated radio-immunoassays for total serum thyroxine and triiodothyronine. Clinica Chimica Acta, 60,25-32.

CRISP, A.H. (1965) Clinical and therapeutic aspects of anorexia nervosa-a study of 30 cases. Journal o f Psychosomatic Research, 9, 67-78.

CROXSON, M.S. & IBBERTSON, H.K. (1977) Low serum T3 levels and hypothyroidism in anorexia nervosa. Journal of Clinical Endocrinology and Metabolism, 44,167-1 14.

DARNFORTH, E., Jr, SIMS, E.A.H., HORTON, E.S. & GOLDMAN, R.F. (1975) Correlation of serum triiodothyronine concentrations (T3) with dietary composition, gain in weight and thermogenesis in man. Diabetes, 24,406, abstract 54.

FEIGHNER, S.P., ROBINS, E., GUZE, S.B., WOODRUFF, R.A. Jr, WINOKUR, G. & MUNOZ, R. (1972) Diagnostic criteria for use in psychiatric research. Archives of General Psychiatry, 26,57-63.

HALL, R., AMOS, J. & ORMSTON, N.J. (1971) Radioimmunoassay of human serum thyrotrophin. British Medical Journal, i, 528-585.

LUNDBERG, P.O., WALINDER, J., WERNER, I. & WIDE, L. (1970) Effects of TRH on plasma levels of TSH, FSH, LH and GH in anorexia nervosa. European Journal of Clinical Investigation, 2,150- 153.

MILLS, I.H., WILSON, R.J., EDEN, M.A.M. & LINES, J.G. (1973) Endocrine and Social Factors in Self Starvation amenorrhoea. In Symposium-Anorexia Nervosn and Obesity. Royal College of Physicians of Edinburgh, no. 42.

MIYAI, K., YAMAMOTO, T., AZUKIZAWA, M., ISHIBASHI, K. & KUMAHARA, Y. (1975) Serum thyroid hormones and thyrotrophin in anorexia nervosa. Journal of Clinical Endocrinology and Metabolism, 40,334-338.

MOSHANG, T. Jr, PARKS, J.S., BAKER, L., VAIDYA, V., UTIGER, R.D., BONGIOVANNI, A.M. & SYNDER, P.J. (1975) Low serum triiodothyronine in patients with anorexia nervosa. Journal of Clinical Endocrinology and Metabolism, 40,470-473.

ROWLAND, C.V. Jr (1970) Anorexia nervosa. A survey of the literature and a review of 30 cases. Inter- national Psychiatric Clinics, 7, 37-137.