Methods and results

We recruited 49 patients with overt hypothyroidism (TSH >20 mU/l, free thyroxine <8.0 pmol/l; mean age 56.6 (SD 12.1) years) from a cohort of female patients followed prospectively in the thyroid research unit of the endocrine outpatient clinic. The patients had underlying thyroid disorders of chronic autoimmune thyroiditis (n=30), treated Graves' hyperthyroidism (radioiodine or surgery; 16), thyroidectomy for simple goitre (2), and treated toxic adenoma (1). We excluded non-thyroidal illnesses in all women. In addition to measuring TSH and peripheral thyroid hormones (free thyroxine and triiodothyronine) we assessed clinical and metabolic markers of hypothyroidism (clinical score, ankle reflex time, creatine kinase, total cholesterol) to evaluate thyroid hormone action at the tissue level.1^–3 To estimate the association between thyroid hormones and markers of tissue hypothyroidism we correlated TSH and thyroid hormone concentrations with the different tissue parameters. The ethics committee for human studies approved the study.

The table summarises the correlation analyses of thyroid hormone concentrations with different markers of tissue hypothyroidism. In contrast to the good correlations with both circulating thyroid hormones, we found no correlation or only weak correlations with serum TSH.

We then used the Mann-Whitney U test to analyse patients' data in relation to the severity of tissue hypothyroidism as assessed by ankle reflex time (patients with moderate prolongation (410–550 ms; n=15) compared with patients with severe prolongation (>550 ms; 15)). Free thyroxine concentrations declined significantly between these groups (from median 5.5 (range 1.9-7.8) pmol/l to 2.5 (1.8-3.5) pmol/l; P=0.001), as did triiodothyronine (1.0 (0.7-1.6) nmol/l to 0.8 (0.2-1.1) nmol/l; P=0.002). We also found an impairment of the clinical score (from 5 (3–11) points to 8.5 (6–11) points; P=0.005), creatine kinase (144.0 (62.0-362.0) U/l to 566.0 (229.0-2170) U/l; P<0.001), and total cholesterol (7.63 (5.7-11.2) mmol/l to 9.4 (7.8-14.2) mmol/l; P=0.003). In contrast, we found no difference in TSH concentrations between the groups (from 42.0 (26.1-137.0) mU/l to 53.8 (23.6-95.3) mU/l; P=0.44).

Comment

TSH is a poor measure for estimating the clinical and metabolic severity of primary overt thyroid failure. This is in sharp contrast to the high diagnostic accuracy of TSH measurement for early diagnosis of hypothyroidism.

We found no correlations between the different parameters of target tissues and serum TSH. Our findings are in accordance with a cross sectional study showing only a modest correlation between TSH and the percentage of positive hypothyroid symptoms4 and data showing discordant responses between the pituitary and peripheral target tissues in patients treated with L-triiodothyronine.5 We assume that secretion of TSH is driven by maximal stimulation, with no further increase occurring with greater severity of hypothyroidism. Therefore, the biological effects of thyroid hormones at the peripheral tissues—and not TSH concentrations—reflect the clinical severity of hypothyroidism. A judicious initiation of thyroxine treatment should be guided by clinical and metabolic presentation and thyroid hormone concentrations (free thyroxine) and not by serum TSH concentrations.