Free thyroxine (FT4) quantification is continuing to be a concern. The purpose of the following study was to evaluate the analytical performance of Tosoh AIA900 based on a one-step technique and its comparison to Access 2 (two-step technique) over different clinical contexts (euthyroid, thyroid disorders, uncontrolled diabetes, renal failure and pregnancy). The protocol established by the French society of Clinical Biology was used to evaluate: imprecision, limit of detection, trueness, linearity, interferences and method comparisons. Within-run variation of 3.1%, 5.7% and 4.4% were found for the low, medium and high controls, respectively. Between-run was 5.8% for low control, 5.7% for medium control and 7.1% for high control. Common interferences did not affect one-step immunoassay FT4 results. The linearity was checked up to 86pmol/L. The limit of detection was 5.5pmol/L. The concordance correlation coefficient (CCC) showed a low agreement (0.6) between both methods. Bland-Altman plot revealed that AIA 900 one-step immunoassay technique provides a significant higher values ((+2.8±2.7pmol/L;p<0.0001). The Passing-Bablok regression demonstrated both proportional and systematic differences in comparison to Access 2. The lowest association was noted in subjects with impaired renal function (CCC=0.27). At the time of the study, the results of on-step immunoassay are not directly comparable with Access 2.