Anabolic-androgenic steroids

The mechanism is a nuclear receptor doing what nuclear receptors do. An androgen crosses the cell membrane, binds the androgen receptor in the cytoplasm, and the receptor sheds its chaperone proteins, dimerises, moves into the nucleus and binds androgen response elements in DNA. There it recruits coactivators and changes transcription. The downstream consequence in muscle is increased protein synthesis and improved nitrogen retention, so more protein is laid down than is broken off ^[1].

There is a second and slower mechanism that explains why the effects persist. Androgens act on muscle satellite cells, the resident stem cells of skeletal muscle. Biopsy work in men given graded doses of testosterone found that muscle fibre cross-sectional area increased, that both type I and type II fibres hypertrophied, and that myonuclear number rose alongside it ^[5]. A muscle fibre is limited in how much cytoplasm each nucleus can support, so adding nuclei raises the ceiling on how large the fibre can become. Those nuclei are not obviously lost when androgen exposure ends, which is the most plausible physiological basis for the observation that former users retain an advantage.

A third contribution is anti-catabolic. Androgens antagonise glucocorticoid signalling in muscle, blunting the tissue-breakdown effect of cortisol, which is part of why recovery between training sessions improves ^[1].

The response scales with the amount of androgen present. In a controlled study that clamped the endogenous axis and then added graded testosterone, the changes in fat-free mass, muscle volume and strength were dose-dependent and linear across the range tested, and so were the adverse changes in haematocrit and in lipids ^[4]. That is the most important structural fact about this class: the benefit and the harm scale together on the same axis, because they are the same receptor being activated in different tissues.

This is the same receptor the body's own testosterone uses. AAS do not add a new mechanism. They deliver far more of an existing signal than the body would ever produce.

Every steroid is described with an anabolic to androgenic ratio, usually quoted against testosterone at 100:100. The intent is intuitive: a higher first number is supposed to mean more muscle per unit of acne, hair loss and prostate growth.

It is worth knowing where those numbers come from. They are almost all derived from a rodent assay in which a compound is given to a castrated rat and two tissues are weighed: the levator ani muscle for the anabolic score and the ventral prostate and seminal vesicles for the androgenic score. The ratio is the quotient. That assay is a real measurement and it was genuinely useful for ranking compounds during discovery, but it was done in rats, on two tissues, over short exposures, and it does not predict what happens in a human across a dozen tissues over months ^[1].

The chemistry that actually produces separation is more specific, and there are three main routes. Resistance to 5-alpha reductase: testosterone is converted by that enzyme into dihydrotestosterone, a several-fold more potent androgen, and the enzyme is concentrated in skin, hair follicle and prostate rather than in muscle. A molecule that cannot be converted there loses the local amplifier, which is roughly the story of nandrolone, where 5-alpha reduction produces a weaker metabolite rather than a stronger one. Resistance to aromatase: aromatase converts testosterone into oestradiol, and a molecule it cannot act on produces no oestrogenic effects. Altered receptor conformation: a different ligand shape recruits a different coregulator set, which is the same mechanism the SARMs were built to exploit ^[15].

The separation is never complete. Every AAS activates the same receptor, that receptor is expressed in prostate, skin, hair follicle, liver, kidney, heart and brain as well as in muscle, and no chemistry so far has produced a molecule that reaches muscle and nowhere else. That limitation is precisely the problem SARMs were invented to solve, and after roughly twenty-five years of that programme there is still no approved product ^[15].

One consequence catches people out. Removing aromatisation removes oestrogenic side effects but oestrogen is not purely a liability in men: it contributes to bone density, to lipid handling and to libido. A compound or an aromatase blocker that drives oestradiol very low produces its own set of problems, and this is a real clinical pattern rather than a theoretical one ^[6].

Testosterone taken by mouth in its natural form is almost useless, because the liver clears nearly all of it on the first pass. Two chemical solutions exist, and each creates its own problem.

Injectables are usually esters. A fatty acid chain is attached at the 17-beta hydroxyl, which makes the molecule far more lipid-soluble, so an oil depot in muscle releases it slowly while enzymes cleave the ester off to free the active hormone. The length of that fatty acid sets the release rate and therefore the injection interval; see pharmacokinetics for the general principle. The ester itself is not pharmacologically active and contributes nothing but timing.

Most oral steroids are 17-alpha-alkylated. Adding a methyl or ethyl group at carbon 17 blocks the liver's main route of inactivation, so a useful fraction survives to reach circulation. The cost is that the same modification makes the molecule persistently hepatotoxic, and this is the best-established organ-specific harm in the entire category. The pattern is typically cholestatic, with jaundice and marked rises in bilirubin; rarer and more serious outcomes include peliosis hepatis, in which blood-filled cavities form in the liver, and hepatic tumours ^[1][6]. This is why liver enzymes are the standard monitoring in anyone using orals, and why combining two 17-alpha-alkylated compounds multiplies the load rather than adding to it. That combination is one of the flags the interactions and stacks engine raises.

A useful diagnostic detail: the enzyme rises seen with orals are not always liver injury. Intense resistance training raises AST and ALT from muscle, so a raised ALT with a normal gamma-glutamyl transferase often reflects muscle rather than liver. Distinguishing them requires more than one marker, which is a reason routine self-interpretation of a single number goes wrong in both directions.

The table below sets out the structural families, what each modification does, and what it costs.

The harm literature is uneven in a specific and important way, and understanding that shape is more useful than any single number. There are almost no randomised trials at the amounts people actually use, for obvious ethical reasons, so the evidence is a mixture of controlled trials at medical doses, cross-sectional studies comparing users with non-users, registry and cohort data, and case reports. That means the direction of most effects is well established while the size of them is not ^[6][2].

The clearest signal is cardiovascular, and it is the one to take most seriously. A controlled study using cardiac imaging in weightlifters found that long-term AAS users had reduced left ventricular systolic function and impaired diastolic function compared with non-using weightlifters, and that coronary artery plaque volume was substantially greater and tracked with lifetime duration of use ^[7]. That is a dose-response relationship on a hard structural endpoint, which is the strongest form of observational evidence available here.

The mechanisms behind it are individually well characterised: a sharp fall in HDL cholesterol with a rise in LDL, particularly with oral compounds; increased haematocrit and blood viscosity; raised blood pressure; and direct effects on cardiac muscle, since cardiomyocytes express the androgen receptor and hypertrophy in response ^[1][6].

The table sets out the organ systems and the strength of the evidence for each.

The most predictable consequence deserves its own section, because it is the one most often described as temporary and most often is not.

The hypothalamic-pituitary-gonadal axis is a thermostat. The hypothalamus releases GnRH in pulses, the pituitary responds with LH and FSH, LH drives testosterone production in the Leydig cells of the testis and FSH supports sperm production, and circulating androgen and oestradiol feed back to turn the whole thing down. An external androgen is read by that feedback loop as an abundance signal, so GnRH, LH and FSH fall, and the testis stops producing. Testicular volume falls, and because sperm production depends on very high local testosterone concentrations that only intratesticular production can provide, spermatogenesis falls with it, often to zero. A meta-analysis across athletes and recreational users found this pattern consistently ^[9].

Recovery is where the honest picture diverges from the common one. A case-control study of former users, no longer taking anything, found significantly lower serum testosterone and a higher burden of hypogonadal symptoms years after they had stopped, compared with matched non-users ^[8]. Anabolic-steroid-induced hypogonadism is a recognised clinical diagnosis with its own management literature, which exists precisely because spontaneous recovery is not universal ^[10]. Longer exposures, higher amounts and older age all make it less likely, and there is no reliable way to predict in advance who will recover fully.

Fertility follows the same pattern and on a slower clock. Sperm production takes roughly two to three months per cycle even once the signal returns, so recovery is measured in months at best, and a minority do not return to baseline. This is a routine reason for referral to fertility clinics ^[10].

Two further points belong here. The first is that the recovery problem is invisible while it is developing, because exogenous androgen masks every symptom of the deficiency it is causing. The second is that this is the single most common source of long-term regret in the surveyed populations, and it is the one that a short exposure at a young age can still produce.

It is worth being explicit about the limits of the evidence rather than letting the confident tone of most writing on this subject stand in for certainty.

Almost nothing is randomised at the amounts actually used. The controlled trials sit at or modestly above replacement levels; non-medical use frequently runs at many multiples of that, in combinations, for months at a time. Extrapolating from one to the other is an assumption, and the linear dose-response seen at the lower end ^[4] is a reason for concern rather than reassurance.

The user populations studied are not random samples. People who use these compounds differ from those who do not in training history, other substance use, risk tolerance and body-image psychology, and no observational study fully removes that. This cuts both ways: it can inflate an apparent harm, and it can hide one.

The long-term cardiovascular picture is still forming. The cohorts are young by cardiology standards. Widespread non-medical use began in the 1980s, so the first large group of long-term users is only now reaching the age at which cardiovascular events become common, and the epidemiology of that is still being written ^[13].

The product itself is unknown. These are black-market goods with no analytical oversight, and mislabelling, wrong compounds, wrong quantities and non-sterile preparation are all routine. Every risk on this page is being taken on top of not knowing what is in the vial.

Which is why this page stops here. Explaining what these drugs do to a body, and what the evidence says the costs are, is information a person is entitled to have. Telling someone how much to take, in what combination, for how long, or where to buy it, is a different act entirely, and this site does not do it. Prescription-grade hormones belong with a physician who can measure what is happening. This is educational information, not medical or performance advice.