Södergård · Journal of steroid biochemistry 1982 · mathematical modeling and observational validation study · n=27

Calculation of free and bound fractions of testosterone and estradiol-17 beta to human plasma proteins at body temperature.

Cited 1102 times in the scientific literature.

Level 5 - mechanism / opinion, no new human data

Bench laboratory measurement and mathematical modeling without a clinical trial design.

PubMed 7202083 · doi:10.1016/0022-4731(82)90038-3 · record verified 2026-08-27

What was done

The authors developed and evaluated a mathematical model to calculate free, albumin-bound, and testosterone-estradiol-binding globulin (TeBG/SHBG)-bound fractions of testosterone and estradiol in human plasma at 37 °C. The model incorporated measured total steroid concentrations, albumin concentration, TeBG capacity, and measured or literature-derived association constants for testosterone, estradiol, 5α-dihydrotestosterone (DHT), 5-androstene-3β,17β-diol (Ae), and 5α-androstane-3α,17β-diol (Aa). Model calculations were applied to plasma samples from 11 normal men and 16 normal women of reproductive age.

What was found

In 11 normal men, calculated mean testosterone fractions were 2.0% unbound, 53–55% albumin-bound, and 43–45% TeBG-bound; estradiol was 2.4% unbound, 68–70% albumin-bound, and 28–30% TeBG-bound. In 16 normal women, testosterone fractions were 1.5% unbound, 36–37% albumin-bound, and 62% TeBG-bound; estradiol was 2.0% unbound, 52% albumin-bound, and 45–46% TeBG-bound. Variations in DHT, Ae, and Aa concentrations did not significantly alter free testosterone or estradiol concentrations, demonstrating these metabolites can be omitted from standard calculations.

Why it matters

This establishes a foundational law-of-mass-action formula for calculating free and protein-bound sex hormone fractions at physiological temperature from standard biochemical measurements.

Limits

The clinical sample was small (n = 27 total) and limited to healthy individuals. Direct physical measurement of free fractions (e.g., equilibrium dialysis) for concurrent validation was not detailed in the abstract, and the model assumes static binding affinities that may vary in pathological states.

Cited by