This 2015 Zurich human review addressed:

“A remarkable lack of discussion on the meaning and interpretation of frequently used hormone ratios.

The interpretation of hormone ratios is complicated and in many cases not sufficiently supported from a theoretical point of view.

Based on the assumption that the balance between two interdependent hormones determines their eventual effects on brain and other tissues, this index has been commonly interpreted as an indicator of the balance between two endocrine systems.

The ratio is typically calculated by simply dividing the raw value of one hormone by the raw value of a second hormone. However, endocrine parameters may fluctuate considerably within individuals across short periods of time on the basis of circadian rhythms or contextual factors. Nevertheless, the ratio method has so far only rarely been applied in the context of repeated endocrine assessments.”

The researchers made a non-exhaustive list of three dozen studies that used hormone ratios among cortisol, dehydroepiandrosterone sulfate (DHEA-S), estradiol, progesterone, testosterone, triiodothyronine (T3), thyroxine (T4), etc., to explain various outcome measures such as:

*“H**ealth status**Aggressive behavior**Psychopathy**Marital violence**fMRI response to angry and happy faces**Early life adversity**Depression**Chronic stress**Alexithymia”*

The researchers’ 2015 study on *“endocrine correlates of pro-environmental behavior”* was used as an illustrative example. The examined study had 229 male subjects between the ages of 19 and 77. Salivary cortisol (C) and testosterone (T) was sampled with these results:

“T/C and C/T ratios produce different means, standard deviations..and distributional properties..which significantly deviate from normality.

Height is not significantly associated with either T/C or C/T. In fact, looking at the original variables, C correlates positively with height while T shows no association.

When we include age as a covariate (assuming that it is associated with both height and hormone status) the partial correlation between T/C and height then is significant while the association between C/T and height is non-significant, even though both ratios are based on the exact same data.

Looking at the negative association between age and T/C the observed age-related ratio decline is mainly due to the fact that the T value in the numerator decreases with age while the C value in the denominator remains relatively constant. In this case, the analysis of the individual variables therefore offers more information and a more accurate picture of the underlying relationships.

A few previous studies have standardized the two underlying hormone distributions before calculating the ratio in order to account for the fact that two hormones often exhibit very different means and standard deviations. Standardization leads to values that express each subject’s hormone concentration relative to the sample mean.

A ratio calculated on the basis of such standardized hormones takes on a different meaning. In particular, the ratio no longer merely represents the proportion of the two hormones within the individual but also incorporates how high the two hormone concentrations are with respect to the sample distributions.”

Practices to improve the use and interpretation of hormone measurements included:

“Regression techniques employed on the original variables constitute a better suited alternative devoid of the problems associated with the ratio method. Moderation analysis, in particular, is a useful approach, which often provides more detailed insight into the relationships of interest.

Ratios should either be analyzed with non-parametric techniques, or be log-transformed before parametric statistical methods are applied.”

Set points and variations in an individual’s hormone balances are usually the effects of underlying causes. Hopefully, one day researchers will pay more attention to effectively dealing with ultimate causes as the preferred methods of managing an individual’s health, behavioral, and neurobiological effects.

http://www.psyneuen-journal.com/article/S0306-4530%2815%2900953-1/fulltext “How to use and interpret hormone ratios”