Diagnosis
  An increased level of free T3 and T4 in combination with non-suppressed TSH in serum is indicative of RTH, but is also observed in TSH-secreting pituitary adenomas. There are however a number of other conditions which can give rise to spuriously elevated T3 and T4 levels with normal TSH. It is therefore important to first rule out any of these other possibilities before embarking on the path to RTH. Of course a careful check of the history will exclude causes like drugs (amiodarone, iodine-containing X-ray contrast agents) and non- thyroidal illness which often are associated with a high serum FT4 but low FT3. Familial dysalbuminemic hyperthyroxinemia gives rise to markedly elevated total T4 but normal FT4 levels in serum. Endogenous anti T3 and T4 antibodies or heterophilic anti-TSH antibodies in the serum cause spurious results; a simple test to rule out the presence of such antibodies is diluting the serum and checking that the level of the analyte measured decreases linearly with the dilution steps.

Having ascertained the validity of the obtained hormone test results, a distinction must be made between a TSH-secreting pituitary adenoma and RTH. This can be difficult when imaging of the pituitary does not show a tumor. In both cases TSH is refractory to thyroid hormone feedback. A TRH test can be helpful. In the case of RTH there will be a response of TSH to the TRH which will be less so when an autonomous pituitary tumor is present. Furthermore the a-subunit to TSH ratio is normal in RTH whereas it will be elevated in TSH secreting tumors (Beck-Peccoz, 1996; Weiss, 1999a).

When the differential diagnosis clearly points at RTH, the patients genomic DNA can be sequenced, in particular exons 5-10, to establish the diagnosis. In our laboratory we use denaturing HPLC (Jones, 1999) to quickly screen for the position of a possible mutation, which can then be sequenced.

In cases where the diagnosis is not clear or when no mutation is found in the receptor gene it is clear that tests aimed at measuring the effect of T3 in peripheral tissues are required. To this end the scheme which was developed by Refetoff and coworkers (Refetoff, 1993) can be used (Fig. 4). It consists of administering increasing doses of T3 (50, 100 and 200 mg/day, each given for three day periods). At the end of each three day period various T3 dependent peripheral tissue function tests are done (table 2).

Fig. 4: Model for gene activation by the TR



When the TR is unliganded it binds to the TRE of the gene as a homodimer (two identical TR monomers) and it represses gene expression by binding to a co-repressor. When hormone binds to TR, the corepressor is released and the homodimer falls apart. The TR receptor monomer then acquires another heterodimerisation partner RXR (the retinoic X receptor) and then binds again to the TRE as a heterodimer (two different monomers) and attracts a coactivator which will signal to the transcription machinery. As a result the gene is actively transcribed.


Table 2: Relative changes in peripheral tissue function tests upon increasing oral doses of T3 in non-resistant subjects (NR) and in resistant subjects (RTH)

test  100 mg T3/day  200 mg T3/day
NR RTH NR RTH
BMR (% increase) 17±3 3±2 20±6 6±1
body weight (% of basal) 98±0 100±0 97±0 100±0
urinary 24h nitrogen (% of basal) 127±16 93±11 133±11 94±10
cholesterol (% of basal) 74±5 96±3 70±3 83±4
triglycerides (% of basal) 80±9 102±5 76±7 105±11
CPK (% of basal) 68±5 135±23 87±15 107±9
ferritin (% of basal) 120±14 95±7 115±13 75±6
SHBP (nM change) +9±5 -5±6 +17±7 +1±7
TSH (increase in response to TRH) 5±2 29±9 2±0 7±4
PRL (increase in response to TRH) 42±7 94±42 24±5 70±17

Adapted from Weiss and Refetoff (2000). Significant differences between NR and RTH subjects are indicated in bold type.


Using this protocol significant changes in these parameters can be found when comparing RTH patients with non-affected subjects. Some of these are also dependent on the T3 dose. For instance changes in cholesterol, creatine kinase and TSH are at their most informative during the 100 mg period whereas urinary nitrogen, triglycerides and ferritin are at their most informative during the 200 mg period. So when results of the dynamic test are compared to the normal response, a diagnosis of RTH can be made.