Background: The causes of subclinical hyperthyroidism have only been reported from clinical studies. Aim: To determine the prevalence and pathological causes of reduced serum TSH levels in subjects recruited from an epidemiological survey. Material/subjects and methods: Serum TSH was measured in 7954 subjects in the 5th Tromsø study. Subjects with serum TSH<0.50 mlU/l, not using T4, without a previous diagnosis of thyroid disease, without serious concomitant disease, and younger than 80 yr, were invited for a re-examination. If low serum TSH was persistent,thyroid scintigraphy was performed. Results: Among the 4962 subjects that met the inclusion criteria, serum TSH was <0.50 mlU/l in 105 subjects. Twelve subjects had a suppressed serum TSH level (<0.05 mlU/l). Two of these were lost to follow-up, 4 had Graves’ disease, 4 had adenoma, and 2 had multinodular goiter. In the 93 subjects with serum TSH 0.05–0.5 mlU/l, 55 were re-examined, of whom 35 had normalized their serum TSH level. In the remaining 20 subjects, 1 had Graves’ disease, 6 had adenoma (of which 2 were toxic adenomas), 7 had multinodular goiter, and 6 were considered normal. Among the 521 subjects using T4, 70 (13.4%) had a suppressed serum TSH level. Conclusions: Most of the subjects with a suppressed serum TSH level will be on T4 medication. Otherwise, if the suppressed serum TSH level is found by chance, this probably represents a clinically important thyroid pathology. Also, in subjects with a persistently low serum TSH level (0.05–0.5 mlU/l) most will have a pathological thyroid scan.