Stability of testosterone measured in male archival serum samples by two different methods.

OBJECTIVE: To investigate testosterone stability in archival serum samples stored for etiological cancer research, and compare two methods for testosterone measurements. DESIGN AND MEASUREMENTS: Four sets of 130 serum samples were randomly selected from male blood donors, aged 40-49 years at the time of blood draw. The sets had been stored at -25°C for 1 month, 4, 17 and 29 years, respectively, and were analyzed for testosterone, sex hormone binding globuline, follicle stimulating hormone, luteinizing hormone, sodium, albumin and cotinine. Testosterone was measured by two methods, an electrochemiluminescence immunoassay, and a liquid chromatography tandem mass spectrometry method. RESULTS: The mean level of testosterone in the samples with the longest storage time (29 years) was substantially higher than that of the fresh samples. The two techniques gave approximately equal results for testosterone values in the range 5-27.5 nmol/L, close to normal range. CONCLUSIONS: The high mean levels of testosterone in the oldest samples suggest a downward trend over the last three decades, as any degradation during storage would tend to give the opposite result. For archival serum samples, both electrochemiluminescence immunoassay and liquid chromatography tandem mass spectrometry are applicable methods for measurement of testosterone within the expected reference range.