Endocrine Abstracts (2009) 20 OC3.4

Reference ranges for sex hormone-binding globulin and free testosterone index in adult men

Nele Friedrich2,3, Henry Völzke2, Alexander Krebs1, Matthias Nauck1 & Henri Wallaschofski3


1Institute of Clinical Chemistry and Laboratory Medicine, Greifswald, Germany; 2Institute for Community Medicine, Greifswald, Germany; 3Department of Internal Medicine A, Greifswald, Germany.


Objective: The majority of circulating testosterone is bound to sex hormone-binding globulin (SHBG), but also to albumin and cortisol-binding globulin. The remaining part is free-circulating testosterone unattached to serum proteins, which represents the active form of the hormone. A common measurement of the free testosterone is the calculated free testosterone index (FTI)=100*(total testosterone/SHBG). Testosterone is the principal male sex hormone and is involved in the regulation of fertility, libido, and muscle mass. The objective of the present study was to calculate age-specific reference ranges for serum SHBG and free testosterone index (FTI) using quantile regression.

Methods: From the cross-sectional population-based Study of Health in Pomerania (SHIP) 806 healthy men were included in the analyses. Serum testosterone and SHBG levels were measured using a competitive chemiluminescent enzyme immunoassay on an Immulite 2500 analyzer. All data were weighted to adjust for non-response and reflect age–sex distribution of the European adult population.

Results: The use of quantile regression provided exact reference ranges. Nearly nearly 5.0% of subjects (equal for SHBG and FTI: above 2.3%; below 2.4%) with SHBG levels as well as FTI values outside the reference range were detected.

Conclusion: The present study established age-specific reference ranges for serum SHBG and FTI levels. Previous studies of our research group (1, 2) regarding reference ranges showed that quantile regression should be preferred to calculate reference ranges because a better concordance to original data is possible due to no distribution assumption are required and the robustness against outliers. These results confirmed this suggestion.

Article tools

My recent searches

No recent searches.