Reference levels for serum thyroid function tests of diagnostic and prognostic significance.

A controversy exists on the value of upper thyrotropin (TSH) reference level. Currently available studies are based on cross-sectional data leaving uncertainty about the prognostic significance of the upper TSH reference level. With the present study we sought to establish reference values for serum thyroid function tests that are of both diagnostic and prognostic significance. We used data from the prospective population-based Study of Health in Pomerania (SHIP) with a 5 year follow-up (6080 person-years). We included data from 1203 subjects (525 women) without prevalent subclinical or manifest thyroid disorders. An event-free reference population was separated comprising 1053 subjects (473 women). When comparing reference values as analyzed from either the whole reference population or the event-free reference population, we observed notable differences in TSH reference intervals. While the lower TSH reference values were similar in both populations, the upper value was 1.95 mIU/l and thus by 7.6% lower in subjects without incident events compared to the whole reference population. Both populations did not substantially differ with respect to serum FT3 and FT4 reference intervals. The upper TSH reference value is lower than recommended when both diagnostic and prognostic significance are considered in the definition of the TSH reference range.