Total testosterone falls by roughly 0.4 percent a year in men aged 40 to 70. The free fraction, the part your body can actually use, falls by roughly 1.3 percent a year, more than three times faster.
That gap is how a man gets handed a normal total testosterone and is still right that something has changed.
You have already had this argument with yourself. Is this just age, or is something actually wrong. The honest answer is that it is a slope rather than a switch, and that the steepest part of the slope does not show up on the test most men are given.
The two numbers
Total testosterone falls by roughly 0.4 percent a year in men aged 40 to 70, beginning gradually from around age 35. Slow enough that no single year of it is detectable from the inside.
The free fraction is another matter. Most of your testosterone travels bound to carrier proteins and is unavailable. The free portion is the part that does the work, and it falls by roughly 1.3 percent a year, more than three times faster than the total.
Which is how a man gets handed a total testosterone in the normal range, gets told he is fine, and is still right that something changed.
How the system is supposed to run
Testosterone is not produced by one gland acting alone. It is the output of a loop between the brain and the testes, the hypothalamic-pituitary-gonadal axis.
The hypothalamus signals the pituitary. The pituitary releases two hormones, LH and FSH. LH reaches the Leydig cells, the small testosterone factories sitting between the sperm-producing tubes in the testes, and tells them to produce. Circulating testosterone then reports back to the brain, which adjusts the next order. It is a thermostat, and like any thermostat it can fail at the sensor, at the controller, or at the furnace.
With age, both ends drift. The signaling from above loses precision. The factory below loses efficiency. The second half is where the more interesting biology sits.
Inside the factory
A Leydig cell builds testosterone out of cholesterol. The cholesterol has to be physically carried into the cell's mitochondria, the engine room, by a small crew of transport proteins called StAR, TSPO and VDAC1. That transport step, not the chemistry that follows it, is the bottleneck of the whole process.
With age the delivery crew slows down. Less raw material reaches the production line even while the brain is still placing orders at the old rate. That is the mechanical core of age-related decline: not a shortage of instructions, a shortage of supply at the point of manufacture.
Two further processes are implicated. Cellular stress can jam a key production enzyme. And autophagy, the cell's internal cleaning and recycling routine, slows with age, so damaged components accumulate inside the factory instead of being cleared out.
Where that evidence comes from. The enzyme finding is mouse work: the stress cut testosterone output, and blocking it restored production. That is a plausible mechanism in an animal, not a proven intervention in a man. The autophagy piece is newer and less settled still. Worth knowing, not worth betting on, and any site that presents it as established is running ahead of the data.
Why it is worth knowing your position
Low testosterone in aging men has been associated with a higher risk of diabetes, dementia, cardiovascular disease and death. Associated is the operative word. These are observed relationships in populations and they do not establish direction: illness lowers testosterone at least as reliably as low testosterone accompanies illness. It is a reason to know where you stand, not a reason to panic about where you stand.
Exercise, with the caveat attached
A single session does something real and brief. In a study of 7 older men, average age 70, short bursts of moderate exercise lifted total testosterone by 39 percent and free testosterone by 23 percent. Levels were back to baseline within 4 hours.
Seven men. That is a hint, not a finding, and it deserves to be read as one. The more useful observation is the longer one: a larger study of 202 men, average age 52, followed for about 15 weeks, found a more sustained rise with regular exercise. Set that alongside the standard guidance of roughly 150 minutes of moderate activity a week with resistance work included, and regular training does appear to support healthier levels over time. Better trials would settle it. They have not been run.
Body fat, which is the larger lever
Fat tissue converts testosterone into estrogen, a process called aromatization. The more excess fat you carry, the more of what you produce is converted away from what you produced it for. It is a leak, and it compounds. Lower testosterone makes muscle harder to hold, less muscle makes fat easier to add, more fat widens the leak.
Closing that loop through diet and activity is the best-supported thing a man can do about his own testosterone without anyone's help. It is also the least commercially interesting item in the field, which is roughly why you hear about it least.
What this is, and what it is not
It is biology. A slow, well-described decline in the efficiency of a cell type that has been doing a demanding job for decades. It is not a character flaw, it is not a discipline problem, and it is not evidence that you let yourself go.
It is also not entirely fixed. Part of the slope is age and cannot be argued with. Part of it is weight, sleep, alcohol, illness and medication, and that part answers to intervention. Separating the two requires measurement rather than introspection, which is the one thing an article cannot do for you. This site sells nothing and prescribes nothing, so read that as arithmetic rather than a pitch.
Where ordinary decline stops being ordinary, and how often that actually happens by age band, is set out in how common it is.