Comparison of 4 Free Thyroxine assays — ASN Events

Comparison of 4 Free Thyroxine assays (#363)

Paul F Williams 1 2 , Kris Tan 1 , Lillian Tan 3 , Elaine Uhr 4 , Nimalie J Perera 5 , Elizabeth L Chua 2 5
  1. Endocrinology Laboratory, Royal Prince Alfred Hospital, Sydney, NSW, Australia
  2. Department of Endocrinology, University of Sydney, Sydney, NSW, Australia
  3. Endocrinology Laboratory, Prince of Wales Hospital, kensington, NSW, Australia
  4. Haematology Laboratory, Royal Prince Alfred Hospital, Camperdown, NSW, Australia
  5. Endocrinology Department, Royal Prince Alfred Hospital, Sydney, NSW, Australia


There is no international reference standard for fT4 measurement. Recent attempts to re-standardise fT4 methods to bring them into better alignment have not been able to overcome the variability between current assays.


Fifty-two fT4>30pmol/L samples were compared using 5 commercially available assays –  Axsym (Ax), Architect (A), Beckman Coulter (BC) , Immulite2000 (IM) and Roche Modular (R) analysers. In addition, 380 samples with fT4<30pmol/L were tested on the A and R analysers.


For fT4>30pmol/L, the mean±SD and median varied widely between methods. The Ax and A were closest to each other with mean of 41.8±15.6 and 40±14.7, and median of 41.7 and 42.1 respectively. BC mean was 46.99 ±15.5 and median was 46.7. Mean of IM was 53.6±19.4 and median was 54.3, while R mean was 64.3±28.1 and median was 67.1. Figure 1 shows the comparison between the individual values for each patient in each assay. Two step assays A and BC gave lower values for fT4 than single step assays.

For fT4<30pmol/L samples tested on A vs R, mean±SD was 22.1±3.2 vs 26.1±4.5 and median was 21.7 vs 25.3. This showed that fT4 levels are much higher on the R assay. Despite the increased harmonisation of fT4 at levels <30pmol/L, 16 more patients would be classified as hyperthyroidism based on higher levels on the R assay (Fig 2).


Standardisation of assays and harmonisation of reference ranges are crucial in the diagnosis of thyroid dysfunction. Non-linearity of fT4 values >30pmol/L in certain assays may influence anti-thyroid medication doses with possible risk to patients. In addition, the difference as to whether a patient is hyperthyroid or not may vary depending on the assay used for fT4.

Figure1                                                                                                 Figure 2