Most supplement protocols borrow pharmacology — and most borrow it wrong.
Open any consumer supplement label and you will find a familiar phrase: "Take three capsules per day for the first week, then one capsule per day thereafter." This is a loading-dose schedule, borrowed by analogy from clinical pharmacology, and in the supplement context it is almost always wrong.
A loading dose is a one-off larger dose used to bring a drug rapidly to its steady-state concentration in plasma. The arithmetic is straightforward: the loading dose equals the desired steady-state concentration times the volume of distribution, divided by the bioavailability. It is justified when the maintenance dose alone would take too long to reach efficacy, where "too long" usually means "more than four half-lives." For a drug with a 24-hour half-life, four half-lives is four days — fast enough that for most chronic conditions a loading dose is unnecessary.
Loading doses are genuinely useful for drugs with very long half-lives where rapid onset matters: digoxin (half-life ~36 hours, used acutely in heart failure), amiodarone (half-life weeks, used acutely in arrhythmia), some antibiotics dosed against rapidly dividing organisms. Note the pattern: long half-life, urgent indication, narrow therapeutic window. Almost no supplement satisfies all three.
Most supplements are water-soluble or rapidly metabolised, with half-lives of 4–12 hours. They reach steady state in two days at the maintenance dose. A "loading week" of triple-dose intake delivers no pharmacokinetic benefit; it simply accelerates depletion of the bottle and inflates first-month revenue. Worse, it delivers a transient supraphysiological exposure that the body's homeostatic mechanisms — first-pass metabolism, hepatic upregulation of clearance enzymes — will respond to by clearing the maintenance dose more aggressively than they otherwise would. The loading dose can therefore reduce, not increase, the steady-state exposure achieved by the subsequent maintenance dose.
There are exactly three well-established cases. Creatine monohydrate has an established loading protocol (~20 g/day for 5–7 days) that reliably saturates muscle phosphocreatine stores 1–2 weeks faster than maintenance dosing alone. Vitamin D in profound deficiency (25(OH)D below 20 nmol/L) is often loaded at 50,000 IU weekly for 6–8 weeks before transitioning to a maintenance dose. Iron in iron-deficiency anaemia is loaded — though oral iron loading is increasingly disfavoured in adults given evidence that intermittent dosing preserves absorption. Outside these three cases, a loading dose on a supplement label is, almost without exception, marketing.
What to do instead
Pick a maintenance dose grounded in the human trial literature. Take it daily for at least four weeks before judging effect. If the product label insists on a loading week, halve the bottle and double the duration; you will end up with the same clinical exposure at half the price.
Berberine is widely sold with a loading-week protocol. Berberine has a plasma half-life of about 4 hours and reaches steady state within 24 hours of three-times-daily dosing. The loading week is pharmacokinetically meaningless; the gastrointestinal side effects it produces are the leading cause of patients abandoning the supplement before realising any benefit. We recommend against any loading schedule on berberine.
Ashwagandha extracts are sold with loading protocols of varying ambition. The major active withanolides have half-lives in the range of 6–8 hours; the cortisol-modulating effect demonstrated in the Chandrasekhar 2012 trial took 60 days at a 600 mg/day maintenance dose to manifest. There is no published evidence that loading shortens this onset. The loading-week instruction on most ashwagandha labels is borrowed by analogy from caffeine and creatine, which are not analogous compounds.
Eight NMN products, one lab
We bought and assayed eight leading NMN products. Two failed identity; three failed dose.
What the Information Theory of Aging really claims — and what it doesn't
The press made it sound like we'd reversed aging in mice. We didn't. But we did show something narrower and more useful — that epigenetic age can be uncoupled from calendar age. A careful, paragraph-by-paragraph reading of Lu et al., Nature Aging, 2023.