Key takeaways

  • DHEA is a precursor, not a signal — tissues convert it inside the cell into testosterone or estradiol for local use.
  • That intracrine route never appears on a serum test, which is why DHEA matters proportionally more in women.
  • It falls about 10% per decade in everyone, which describes ageing rather than proving a cause of its symptoms.
  • "Pregnenolone steal" is incorrect: cortisol and DHEA are made in separate adrenal zones with no shared pool.
  • Randomised evidence is real in adrenal insufficiency and much weaker in healthy age-related decline.

DHEA is the most abundant steroid hormone in human circulation and the one most people have never had measured. It is not a hormone in the ordinary sense of a molecule with a target and an effect. It is stock — a circulating reserve of raw material that tissues draw on and convert into whatever they need locally. That makes it different from testosterone or estradiol both to think about and to read on a lab report, and it is why the honest account of DHEA is more cautious than the one usually sold with it.

What DHEA is

Dehydroepiandrosterone is produced almost entirely by the zona reticularis of the adrenal cortex, with a small gonadal contribution (Miller & Auchus, Endocr Rev 2011). Most of it is immediately sulfated to DHEA-S, and that sulfated form is what circulates in bulk.

The sulfation is not a detail. DHEA-S circulates at concentrations several hundred times higher than free DHEA and with a far longer half-life, which makes it a stable reservoir rather than a signal — and is why it is the form measured on a blood test. Free DHEA follows the adrenal's diurnal rhythm, so a single measurement of it tells you little.

DHEA itself has only weak direct androgenic activity. Its significance is as a precursor: tissues expressing the relevant enzymes take it up and convert it, inside the cell, into testosterone or estradiol for local use. This is intracrinology, and it is the key idea here — a meaningful share of a woman's tissue-level androgen and estrogen exposure is manufactured this way and never appears in a serum measurement at all.

Where it sits in the cascade

The pathway branches early and repeatedly from one starting molecule:

Conversion of DHEA onward to testosterone happens in peripheral tissue rather than in the adrenal. In women that route contributes a substantial share of total androgen exposure, which is why DHEA matters proportionally more in women than in men, where testicular production dominates. A man with low DHEA-S and normal testosterone has a much smaller problem than a woman with the same finding.

The decline, and what it does not mean

DHEA-S peaks in the mid-twenties and falls by roughly 10% per decade thereafter (Orentreich et al., J Clin Endocrinol Metab 1984). Typical figures:

This is one of the most consistent age-related endocrine changes there is, sometimes called adrenopause — and it is where the reasoning most often goes wrong. A decline that is universal and predictable is not obviously a pathology, and the fact that DHEA-S correlates with health in older adults does not establish that the low value is causing the poor health. Chronic illness, inflammation and poor nutrition all lower DHEA-S themselves, so a low result in a sick person may be a marker of being sick rather than a reason for it. That makes it a reasonable prompt to examine the rest of the adrenal and gonadal picture, and not a diagnosis on its own. The DHEA-S marker in depth covers interpretation.

Stress, cortisol, and the "pregnenolone steal" claim

Chronic stress is associated with lower DHEA-S and a lower DHEA-to-cortisol ratio. That observation is real; the explanation usually attached to it is not. The popular account is "pregnenolone steal" — under stress the adrenal diverts a shared pregnenolone pool toward cortisol, starving DHEA production. It is an appealing picture and it does not match adrenal physiology. The zona fasciculata makes cortisol and the zona reticularis makes DHEA in anatomically separate cell layers, each with its own enzyme complement, and pregnenolone is generated inside cells from cholesterol rather than drawn from a shared reservoir the two zones compete over. There is no common pool to steal from.

What appears to happen instead is that the zona reticularis is independently downregulated during chronic illness, inflammation and stress while ACTH-driven cortisol output is maintained. The result — cortisol up, DHEA down — looks exactly like diversion without being it. The mechanism is worth stating correctly, because the wrong one leads to the wrong intervention: supplementing pregnenolone to "refill the pool" addresses a bottleneck that does not exist. The HPA axis in depth covers what does drive the pattern.

What low DHEA feels like, and why that is hard to use

The reported symptom set is broad: unexplained fatigue, reduced libido — particularly in women — low mood, a flattened sense of wellbeing, reduced exercise capacity, reduced muscle mass, lower bone density. Every item is also produced by low thyroid, low testosterone, iron deficiency, depression, poor sleep, sleep apnoea and simple under-recovery. None of it is specific, which makes DHEA a poor place to start an investigation and a reasonable place to end one — worth checking after the commoner and more treatable explanations have been excluded, not instead of them. A complete hormone panel puts it in context.

Supplementation, and where the evidence actually is

The strongest evidence is in adrenal insufficiency, where the cortex genuinely is not producing it and replacement has been studied in randomised trials, with reported benefit for wellbeing and sexual function in women (Arlt et al., on DHEA replacement in adrenal insufficiency). In age-related decline in otherwise healthy people the picture is weaker and less consistent, with trials reporting modest or absent effects on the endpoints that matter most (Legrain et al., J Clin Endocrinol Metab 2000). That is the honest summary, and it should shape expectations before anyone starts.

Where it is used, the practical points are:

Two cautions belong here. It is a hormone precursor sold as a supplement, so manufacturing consistency is not guaranteed. And because it converts to both androgen and estrogen, anyone with a hormone-sensitive cancer history should not take it without oncology input. DHEA supplementation for men and testosterone for women cover the two cases in more detail.

What to measure alongside it

DHEA-S is the test to order, read against age-adjusted ranges. The common practice of targeting a young-adult range instead has a rationale rather than an evidence base, and should be described that way. Companion markers that make the result interpretable:

The clinical pearl: DHEA is inventory, not instruction. It matters most in women, in adrenal insufficiency, and in people whose whole steroid profile sits low together — and least as a standalone number in a man with normal testosterone. Treat a low value as a question rather than an answer.

Bottom line

DHEA is an adrenal precursor that circulates in bulk as DHEA-S and is converted inside target tissues into androgens and estrogens for local use — which is why it matters disproportionately in women and why serum levels understate what it does. It declines about 10% per decade in everyone, and that decline describes ageing rather than proving a cause of its symptoms. Replacement has real evidence in adrenal insufficiency and much weaker evidence in healthy age-related decline. The "pregnenolone steal" explanation is popular and incorrect, though the observation behind it is genuine. Measure it as part of a panel, retest the downstream hormones rather than only the precursor, and make any supplementation decision with a physician.

Educational content, not medical advice. Laboratory interpretation and any treatment decision are made by a licensed physician after individual evaluation. Individual results vary.

Inventory
a reserve tissues convert locally, not a direct signal
Women first
where the peripheral conversion route matters most
Steal is a myth
the observation is real; the mechanism usually given is not