Cypionate half-life: eight days or under four?

Two numbers, both real, both sourced Strong evidence

Search for the half-life of testosterone cypionate and you will be told 8 days. Use our calculators and you will be told 3.84. Neither figure was invented. They come from two different kinds of study asking two different questions, so the disagreement is real pharmacology rather than somebody being careless.

Almost nobody says which one they are quoting. That is the actual problem, because the two numbers imply schedules that differ by a factor of two.

Where 8 days comes from Strong evidence

The 8 day figure is the FDA-approved prescribing information for Depo-Testosterone. It is not folklore. It is the number on the label, which is why it propagated everywhere. Worth noting the label states it without saying how it was determined, giving no study and no method. It is consistent with a single-injection terminal-slope estimate, which is what the surrounding text describes.

What rarely travels with it is the sentence next to it. The label explains that testosterone esters in oil injected intramuscularly are absorbed slowly from the lipid phase, so cypionate can be given at intervals of two to four weeks. That is the clinical regimen the label was written for: large infrequent injections for hypogonadism, an approach almost nobody on a modern protocol uses.

So the 8 days is internally consistent with the rest of the label. Lifting it out of that context onto a weekly or twice-weekly schedule is where it stops describing anything real.

Where the four-day figure comes from Strong evidence

The other number comes from the one published population pharmacokinetic model for the ester. Bi and colleagues studied 31 healthy men in a randomized double-blind trial, dosed at 100, 250 or 500 mg per week, across 14 weekly injections. Total testosterone was described by a one-compartment model with first-order absorption, with population estimates of 2.6 kL per day for clearance plus 14.4 kL for volume of distribution. Those are apparent values, which is why the units look like kilolitres. Of the covariates tested, the paper concludes baseline weight is the only one of likely clinical importance.

The half-life the paper reports is 4.05 days, the median of the individually fitted post hoc values across those subjects. That is the figure to quote when citing this study.

Our tools carry 3.84 days, which is a different quantity: it is ln2 times the population volume over the population clearance, so it is a derivation from the paper's parameters rather than a number the paper prints. The two sit less than a day apart in plateau terms, so nothing practical turns on the choice, but they should not be confused with each other. If you are citing Bi, cite 4.05.

Why they disagree: the tap and the drain Strong evidence

Picture measuring how fast a bath drains. There is a drain, which is your clearance, plus a tap dripping into the tub, which is your own testosterone production.

Ignore the tap and the water level falls more slowly than draining alone can explain. You conclude the drain is slow. Turn the tap off and the level now falls at exactly the true drain rate.

Turning the tap off does not make the drain slower. It makes what you observe finally match the drain. That single distinction is the whole disagreement between 8 days plus 4 days, so it is worth holding onto.

Bi's team demonstrated it directly. Their model included an explicit endogenous secretion term: a basal rate of 6.24 mg per day plus up to 12.5 mg per day more driven by LH. Even with LH fully suppressed the basal term keeps running, which is precisely why deleting the whole thing forces the half-life up. Then they removed it plus refitted the same blood samples from the same 31 men, in the authors' words: "When we assume endogenous testosterone secretion is 0 in the PPK analysis, the median estimated half-life increases to 6.87 days."

That was not a new experiment. It was the same data with the tap ignored, reproducing almost exactly the 6.9 days that the earlier noncompartmental analysis had reported. The two are not independent groups disagreeing either. That 6.9 figure comes from MacIndoe 1997. Bi 2018 is a re-analysis of that same trial's data, with Perry, who designed and ran it, an author of both. Same men, same blood samples, two methods.

Their own explanation is that the discrepancy results from failure to consider endogenous testosterone production. So 6.9 is best understood as the answer to a different question: what half-life would you wrongly infer if you attributed a man's own testosterone to his injection?

Does the four-day figure apply to someone on TRT? Mixed

This is the objection worth taking seriously, because it sounds decisive: a man on TRT makes little or none of his own testosterone, so surely he is the no-endogenous-production case, so surely 6.87 is his number.

It gets the analogy backwards, in a way that is easy to do. Assuming zero endogenous production in an analysis is not the same as a man having zero endogenous production. In the first case the testosterone is really there in the blood samples plus the model has been told to pretend otherwise, so it stretches the drug's half-life to absorb testosterone it cannot otherwise explain. In the second case there genuinely is no extra testosterone, so there is nothing to misattribute. One is a fitting artifact. The other is a clean curve.

There is also no mechanism by which suppression could slow clearance. Testosterone from a testis plus testosterone from an ampoule are the same molecule, cleared by the same liver at the same rate. Shutting down your own production changes how much testosterone is present. It does not change how fast the body removes it.

The decisive point is in the paper itself. Bi's subjects were suppressed while they were being dosed. The paper notes that the upregulation of LH on endogenous secretion was completely lost from week 9 to week 19 in the 250 mg plus 500 mg groups, plus from week 13 to 17 at 100 mg, against dosing that ran from week 2 to week 15. So the 4.05 estimate is drawn overwhelmingly from men in the state that matters here: on testosterone, own production shut down. The paper adds that the 299 samples used to determine the main PK parameters were drawn within 26 days of the last dose, a window sitting inside that suppression period.

The endogenous term does its real work at baseline plus during recovery, which is exactly why deleting it wrecks the fit. Four days is not a healthy-volunteer number that needs adjusting upward for suppressed men. It is the number from suppressed men, with the periods either side handled properly instead of being smeared into the drug's half-life.

The honest limit is that no study has measured cypionate kinetics directly in suppressed TRT patients. Bi's authors say as much, flagging that the model was built in 31 healthy men and that re-estimation may be needed before applying it to a patient population. They also note patients may have lower baseline endogenous secretion than their estimate, which is the direction that makes a four-day figure fit a suppressed man better rather than worse.

The one place a longer number is right Mixed

There is a phase where a four-day half-life genuinely under-predicts, so it is worth naming rather than glossing over.

Coming off, the axis recovers. Bi tracked this: endogenous secretion returned to baseline at around week 23, against dosing that stopped at week 15, so about eight weeks of recovery. Through that stretch your own production is being added back underneath the exogenous decay, so measured total testosterone falls more slowly than the drug alone would.

Note what that is. It is the tap being turned back on, which is the opposite of no endogenous production rather than an example of it. If you are predicting how quickly total testosterone drops after stopping, expect the real curve to flatten out above what a four-day half-life implies, plus expect that flattening to be your recovery rather than the ester lingering. Which compound governs the wait is a separate question the PCT calculator answers.

What flip-flop does explain Strong evidence

Flip-flop kinetics often gets offered as the explanation for the gap. It is real here, though it explains a different thing, so the two are worth separating.

Testosterone's own elimination is fast. The paper makes the point with a comparison: testosterone undecanoate has a half-life around 150 minutes taken orally, against 21 to 34 days injected intramuscularly. Same molecule, two orders of magnitude apart. That gap can only be the oil depot, so absorption is far slower than elimination plus the slower process governs the curve. Every multi-day figure you will see for an injectable ester is describing the depot releasing rather than the hormone clearing.

So flip-flop explains why these numbers are days instead of minutes. It cannot explain why two sources measuring the same absorption-controlled process differ by a factor of two. Bi's authors raise it as a limitation of their own absorption model rather than as the reason for the discrepancy, noting that having only one or two samples per subject in the absorption phase prevented a more realistic model.

What it changes in practice Mixed

Levels are within about 5% of their plateau by roughly 4.3 half-lives, so that lands near 18 days on the four-day figure against about 35 on the label figure. Clearance at five half-lives is about 20 days against about 40. Those are not academic differences.

The bloodwork row is where it bites hardest. Waiting six weeks after a dose change, on the belief that levels have not settled, costs you three weeks of information about your testosterone level on a protocol that plateaued around eighteen days. Hematocrit plus how you actually feel move on a slower clock, so a later draw is not wasted, it is answering a different question. Drawing at three weeks on the label figure looks premature when it is not. The bloodwork guide covers the rest of the timing rules.

Which number should you use Mixed

Use the one that matches the question you are asking.

Planning a weekly or twice-weekly protocol, working out when levels plateau, deciding when a blood draw becomes informative: the multiple-dose model is the right instrument, so roughly four days. That is what our half-life calculator plus the cycle calculator use, which is why their numbers sit below what you will read elsewhere.

Reading the label's own two-to-four-week regimen, or reasoning about how long one injection keeps releasing: the label figure is internally consistent, so 8 days is the coherent choice there.

For context, enanthate sits at about 4.5 days in the same database, close enough to cypionate that the two are practically interchangeable on schedule, which means there is no good evidence for the common claim that cypionate is the longer ester. Propionate is about 1 day, which is why it cannot be run on the same frequency. The ester calculator covers the mass differences between them, which is a separate question from the timing one.

What we will not do is print a number without saying where it came from. That is the entire reason this page exists.