Hypothyroidism has been defined as those conditions which result in suboptimal circulating levels of thyroid hormones. The problem is how to define such suboptimal levels? In the 1970s, the advent of routine assays for TSH together with the concept that the most sensitive peripheral tissue to diminished circulating thyroid hormone concentrations is the pituitary led to the acceptance that elevated serum TSH was required to define primary hypothyroidism. There are problems, however, in defining what is an elevated serum TSH and what serum concentrations of TSH should trigger thyroxine replacement therapy. Routine TSH assays show considerable between method biases, which do not appear to correlate with the TSH reference ranges suggested by the manufacturers. These assay bias differences are often ignored in research communications that present data supporting clinical action limits for the implementation of T4 therapy. Furthermore it has been suggested that the log normal distribution of TSH observed in the healthy population may arise through the inclusion of subjects with occult hypothyroidism. If this were the case then the upper reference limit for TSH should be significantly lower than that currently used by most laboratories. In this lecture these problems and current evidence will be discussed.