Enzyme Immunoassays May Be Inaccurate: Familiar Correlates of Male Testosterone (T) Replicate in Britons when T is Measured by Mass Spectrometry but Do Not Replicate in Americans When T is Measured by Enzyme Immunoassay

ABSTRACT Several correlations have been reported between men’s testosterone (T) and other characteristics, e.g., T declines during the day, declines with obesity, and declines with advancing age. We asked if these relationships hold in older men when T is assayed from saliva. Here seven familiar correlations are tested among older American men, their salivary T measured by enzyme immunoassay (EIA). Some correlations can also be tested among older British men, their salivary T assayed by liquid chromatography-tandem mass spectroscopy (LC-MS/MS), a technique noted for its specificity. Most of our attempts at replication failed in the American data but succeeded in the British data. We conclude that failure to replicate in Americans is likely due to inaccuracy of EIA values for T, and that most T correlations hold true in older men when salivary T is accurately measured by LC-MS/MS.