Sports · Part Four · Recovery and Fueling
Lesson 48 / 64
Supplements
A short list of evidence-backed compounds moves performance, and the rest is noise an elite athlete cannot afford to ingest.
Sports supplements are compounds taken alongside food to raise performance, and only a short list survives testing: creatine, caffeine, beta-alanine, dietary nitrate and sodium bicarbonate, with glycerol beside them as a hydration agent. Each one is an input with a dose, and it meets an athlete who is either depleted in the system it acts on or already full. The Unified Model of Tone reads that meeting as the reason one compound moves one athlete and does nothing measurable for the next.
Group A performance list
Six compounds
Undeclared steroids found
14.8 percent of 634 products
Caffeine by CYP1A2 genotype
6.8 percent faster or 13.7 percent slower
Nitrate above 65 mL/kg/min
d = 0.02, not significant
Ergogenic aid.
A compound taken to raise work output beyond what training and food already produce. Membership in the category is decided by controlled trials of exercise capacity and performance, never by the claim printed on the container.
Why a dose meets a state.
Each of these compounds fills a store or acts on a receptor that is already partly occupied. The size of the response tracks how far that store sits from full, so the variable being measured is the athlete rather than the powder.
01What the trials and the screens show
The Numbers Behind the Supplement Shelf
Eight findings that separate the compounds with evidence from the bottles with risk.
02The Group A short list
Only a Short List of Supplements Has Trial Evidence Behind It
Most of what is sold to athletes has no performance evidence, and the products that do have it can be counted on two hands. The IOC consensus statement on dietary supplements found that only a few of the products sold for direct performance benefit have good evidence of it Maughan 2018. It named caffeine, creatine, specific buffering agents and nitrate.
The AIS Sports Supplement Framework sorts the same field into four groups and reserves Group A for ingredients with strong scientific evidence for use in specific situations in sport. Six compounds sit in its performance column: caffeine, beta-alanine, dietary nitrate as beetroot juice, sodium bicarbonate, creatine monohydrate and glycerol. Five more sit in a separate medical column: calcium, folate, iron, vitamin D and zinc.
The discipline is subtraction. Supplementation sits at the bottom of the hierarchy, beneath total calories, macronutrients, micronutrients and meal timing, all of which Sports Nutrition carries. Get those four right and most supplements have nothing left to add. Get them wrong and no powder rescues the program.
So the first question is never which supplement, but whether the diet already supplies what the body is missing. Everything else is convenience, marketing or risk. The few compounds that survive that filter are added one at a time, each with a performance question attached, so their effect can be seen rather than assumed. A cabinet full of capsules usually signals a plan that has lost the thread between supplement and outcome.
Where protein powder actually sits
Protein powder is food in a convenient form rather than a true ergogenic aid, and the framework classes it that way. Isolated protein and mixed macronutrient supplements sit in the Group A sports foods column, alongside sports drinks and gels, and not in the performance column with creatine and caffeine. That is a statement about what a powder is, not a demotion of protein.
Whey delivers leucine, the amino acid that triggers the mTOR pathway, in a form that survives a bag and a shaker. The daily target and its distribution across meals belong to Sports Nutrition. Powder is a delivery method, not magic. Whole food reaches the same number, and the powder simply makes the dose convenient when appetite or schedule fails.
03Creatine and caffeine
Creatine and Caffeine Carry the Largest and Best Replicated Effects
Creatine and caffeine are the two proven aids an elite program can defend on evidence alone, and they work on different systems. Creatine expands the phosphocreatine pool that regenerates ATP during short maximal efforts, which is the energy pathway Exercise Physiology describes in full. Caffeine never reaches that pathway. It works on the brain.
The position stand on creatine concludes that supplementation raises intramuscular creatine concentration, and that the rise may help explain improved high intensity exercise performance and the greater training adaptations that follow Kreider 2017. The same review records enhanced post-exercise recovery, which shows up in a program as more repeated high-intensity work and quicker between-set recovery. Its relevance is largest for athletes whose events run on the creatine phosphate system: throwers, sprinters, lifters and anyone repeating maximal efforts with short rest.
What the loading studies actually did
The original uptake work is worth reading in its own units. Harris and colleagues gave 5 grams of creatine monohydrate four to six times a day for two or more days and measured quadriceps muscle directly in 17 subjects Harris 1992. A single 5 gram dose, they noted, matches the creatine content of 1.1 kilograms of fresh uncooked steak.
Uptake was fastest at the start. In three subjects taking six 5 gram doses a day, 32 percent of the administered creatine entered muscle over the first two days. Renal excretion accounted for 40 to 68 percent of the dose over those first days. Roughly a fifth or more of what was taken up appeared as phosphocreatine, and muscle ATP content did not change.
That is the loading protocol every label paraphrases, and it explains why a maintenance dose is small once the store is full. The position stand describes a habitual dietary creatine intake near 3 grams a day. It records supplementation up to 30 grams a day for five years as safe and well tolerated in healthy people Kreider 2017. Efficacy and safety are separate questions, and creatine has a good answer to both.
Caffeine is a central nervous system drug
Caffeine sharpens the nervous system rather than the muscle. It blocks adenosine receptors in the brain. That is why the same dose which steadies a sprinter in the blocks also holds decision speed late in a match, the quality Reaction Time and Motor Control measures. The position stand records consistent improvement at 3 to 6 milligrams per kilogram of bodyweight Guest 2021.
The same document is specific about the edges of that range. The minimum effective dose may be as low as 2 milligrams per kilogram. Very high doses near 9 milligrams per kilogram bring a high incidence of side effects without extra benefit. The most common timing is 60 minutes before exercise. Caffeine also improves attention and vigilance, and it holds performance up under sleep deprivation, which is the case Sleep and the Athlete examines.
Both compounds are dose sensitive and individual. The right amount is found in training, never discovered for the first time on competition day. The IOC consensus makes the same point as a rule: a supplement intended to enhance performance is tested thoroughly in training or simulated competition before it is used in competition Maughan 2018.
04Buffers and blood flow
Beta-Alanine, Bicarbonate and Nitrate Attack the Chemistry of Fatigue
Three of the six Group A performance compounds work on fatigue chemistry rather than on the drive that creatine and caffeine support, and they attack it from different sides. Beta-alanine raises the muscle's internal buffer. Sodium bicarbonate raises the buffer outside the cell. Dietary nitrate works on blood flow and on the oxygen cost of the work itself.
Four weeks of beta-alanine at 4 to 6 grams per day significantly raises muscle carnosine, which acts as an intracellular pH buffer against the acidity that builds during hard efforts Trexler 2015. The gain shows up most clearly in open end-point tasks and time trials lasting one to four minutes, and the position stand also records attenuated neuromuscular fatigue.
The measured size of that gain is small. Pooling 40 studies, 65 exercise protocols and 1,461 participants gives an overall effect size of 0.18 Saunders 2017. Exercise duration moderated the result, and within the half minute to ten minute window the effect was larger for exercise capacity at 0.50 than for performance at 0.11.
The dose is a schedule, not a single act
Beta-alanine needs weeks of consistent loading to build carnosine. A single dose before a race does nothing but sting. The tingling has a name, paresthesia, and it is the only side effect the position stand reports Trexler 2015. Splitting the daily total into 1.6 gram doses or using a sustained release form attenuates it.
Sodium bicarbonate runs on a shorter clock. Doses of 0.2 to 0.5 grams per kilogram improve performance in high intensity work lasting 30 seconds to 12 minutes Grgic 2021. The position stand identifies 0.3 grams per kilogram as the optimal single dose and 60 to 180 minutes as the usual window before the effort. Bloating, nausea, vomiting and abdominal pain are the common side effects, and they can cost more performance than the buffer buys.
The two buffers combine better than either works alone. Beta-alanine taken alongside sodium bicarbonate produced the largest effect in the pooled analysis at 0.43, roughly double the beta-alanine effect on its own Saunders 2017. Two buffers on two sides of the cell membrane, one shared problem.
Dietary nitrate is the third route. Nitrate from beetroot juice converts to nitric oxide, which widens blood vessels and lowers the oxygen cost of submaximal work. Across 80 randomized crossover studies the pooled effect was d = 0.174 Senefeld 2020. That is a real average, and it was built almost entirely on recreationally active young men.
None of these compounds outranks the recovery built by sleep, fueling, and a regulated central integrative state, the autonomic readiness that decides whether the body can express any of this on the day. A buffer cannot fix a nervous system stuck in sympathetic overdrive. It can only add a clean margin once the foundation already holds.
05Why the same compound splits
The Same Supplement Moves One Athlete and Leaves the Next Unchanged
The supplement literature looks inconsistent because the compound is only half of the experiment. The other half is the athlete who swallows it, and specifically how much room is left in the system that compound acts on. Three of the best studies on this page are really one finding told in three chemistries.
Start with nitrate. The pooled effect of d = 0.174 across 80 studies came almost entirely from recreationally active young men Senefeld 2020. In the 26 studies of well trained endurance athletes at or above 65 mL/kg/min the effect was d = 0.021 and did not reach significance. The same analysis found no significant effect in the six studies that enrolled only women, which is a statement about how few trials have enrolled them.
The single trial version is sharper. Eight elite male 1500 meter runners, averaging 80 mL/kg/min with personal bests near 3:56, took beetroot juice for eight days Boorsma 2014. On the test days the drink carried 19.5 millimoles of nitrate in 210 milliliters. Plasma nitrate rose from 37 to 615 micromoles per liter acutely and to 870 with continued dosing. Submaximal oxygen cost did not move at 50, 65 or 80 percent of peak, and the 1500 meter time trial did not move either.
The compound arrived. The plasma measurement proves delivery, so the null is not a dosing failure. What the group did not have was room. Two of the eight runners improved anyway, by 5.8 and 5.0 seconds acutely, which is exactly the pattern a group mean erases.
Deficiency correction is a different question from enhancement
Correcting a deficiency and enhancing a full system are two different interventions that share a shelf. Iron deficiency is diagnosed in roughly 15 to 35 percent of female athlete cohorts and 5 to 11 percent of male ones Sim 2019. Exercise itself raises hepcidin, the master iron regulatory hormone, which is one reason athletes drift low while eating normally.
Iron and vitamin D are corrected from bloodwork, not guesses. The AIS Sports Supplement Framework keeps them in a separate medical column for that reason, alongside calcium, folate and zinc. Nobody expects iron to make a replete athlete faster. Everybody expects it to change what a deficient athlete can do, and the creatine uptake data says the same thing in muscle: the rise was largest where the store started lowest Harris 1992.
The state is not always a store
Sometimes the starting state is an enzyme rather than a store. Among 101 competitive male athletes, caffeine at 4 milligrams per kilogram cut 10 kilometer time trial time by 6.8 percent in athletes carrying the AA variant of CYP1A2 Guest 2018. In CC carriers the same dose made them 13.7 percent slower. The AC group showed no effect at all.
One dose, one task, three directions. The caffeine position stand names the same variation and adds habitual intake to the list of things that shape it Guest 2021. An average effect across that sample would describe none of the three groups accurately.
This is not a rule that elite athletes are immune to supplements. Beta-alanine is the check on that reading, because its pooled effect was not moderated by training status and not moderated by the total amount ingested Saunders 2017. Endurance training optimizes the oxygen cost pathway that nitrate targets. It does not fill the muscle carnosine store that beta-alanine builds, so the room stays open and the compound still lands.
06Contamination and strict liability
A Supplement Can Cost an Athlete Eligibility With Nobody Intending It
Every product an elite athlete takes is verified against anti-doping rules and third-party tested for contamination, without exception. The contamination record is why that rule reads that way. Between October 2000 and November 2001, investigators bought 634 non-hormonal nutritional supplements in 13 countries from 215 suppliers and analyzed them by gas chromatography and mass spectrometry Geyer 2004. Ninety-four of them, 14.8 percent, contained anabolic androgenic steroids that appeared nowhere on the label.
The detail inside that study matters more than the headline number. Products from companies that also sell prohormones were positive 21.1 percent of the time, and products from companies that do not sell them were positive 9.6 percent of the time. Steroid concentrations ran from 0.01 to 190 micrograms per gram. Nandrolone prohormone intake totaling more than 1 microgram produced positive doping results for norandrosterone for several hours.
Two decades later the picture had not improved where the risk is highest. Screening 66 sports nutrition supplements from 21 brands and 17 web shops, an accredited laboratory found undeclared doping substances in 25 of them, or 38 percent Duiven 2021. The list included the stimulants oxilofrine and methylhexaneamine, the anabolic steroid boldione, the beta-2 agonist higenamine and the beta blocker bisoprolol. Used as prescribed, three of the products could impose a general health risk.
None of that is a defense at a hearing. Article 2.1.1 of the World Anti-Doping Code states that it is the athletes' personal duty to ensure that no prohibited substance enters their bodies. Athletes are held responsible for any prohibited substance found in their samples. Intent, fault, negligence and knowing use do not have to be shown for the violation to stand. The Medical-Legal Standard covers how that duty is documented and where scope of practice sits.
What certification does and does not buy
The supplement industry has no reliable standard for what actually ends up in the bottle, so third-party batch testing is the only check an athlete controls. Contaminated or mislabeled products have ended careers, because WADA holds athletes strictly liable for what is in their body. USADA states plainly that no dietary supplement can be guaranteed to be 100 percent risk free, and it currently recognizes NSF Certified for Sport as the program best suited to reducing supplement risk for athletes. Informed Sport runs a comparable batch testing scheme, and the Supplement 411 high risk list catalogs products already known to cause anti-doping problems.
A product without independent testing is a gamble no professional can take. A product with it is a smaller gamble, taken knowingly, on a batch rather than on a brand. The aim is a small, verified stack that adds a clean margin while protecting eligibility and long-term health.
Harm is measured outside the doping question too
Supplement harm shows up in national health data, entirely separately from eligibility. Surveillance across 63 United States emergency departments produced an estimate of 23,005 visits and 2,154 hospitalizations a year for adverse events tied to dietary supplements Geller 2015. Weight loss and energy products accounted for 71.8 percent of the visits involving palpitations, chest pain or a racing heart, and 58.0 percent of those involved people aged 20 to 34.
Purity is one half of the calculus, and honesty about benefit is the other. Creatine and caffeine have the evidence, and they still have to arrive in a clean container. Many marketed compounds show no measurable effect, and high dose antioxidants taken around training can blunt the adaptation the training was meant to produce, which Inflammation and Healing covers. Anabolic agents and prohormones sit entirely outside legitimate sport and are not a supplement question at all.
Chiropractic and neurological care introduce no substance into the body. That makes this practice drug free and anti-doping compliant by construction, with no batch to verify and no contamination exposure to manage. It is a fact about the modality rather than a ranking of it against any other form of care.
07What we corrected
Six Corrections to the Supplement Claims This Page Used to Make
This page previously counted five supplements with strong evidence. The AIS Sports Supplement Framework lists six in its performance column, and the sixth is glycerol, a hyperhydration agent whose protocols belong with fluid strategy in Hydration and Electrolytes. The count is now six.
The beta-alanine dose stated here was 3 to 6 grams daily. The position stand concludes 4 to 6 grams daily for at least four weeks Trexler 2015, and that is the figure the page now carries. The creatine loading protocol is likewise stated in the terms the original uptake trials used rather than as a rounded daily total.
The protein targets that once occupied a section here have moved to Sports Nutrition, which owns intake guidelines for the section. The claim that nitrate is timed two to three hours before effort is not carried, because it could not be traced to any source cited on this page.
A gold pull-quote attributed to Dr. Jason Dulberg once sat between the sections above. He did not say or write those words, so they are gone. What is left on the page is either sourced to a named trial or labeled as the model's own claim.
08The model's claim
A Supplement Response Is a Reading of the State It Met
Sort the evidence above into two piles and the contradictions dissolve. The first pile is measurement: micromoles of plasma nitrate, percentage points off a 10 kilometer time trial, effect sizes with their confidence intervals, and 94 contaminated bottles out of 634. Every number in that pile belongs to the investigators named beside it.
The second pile holds the question those numbers keep raising and never answer. Why did one input produce a gain in one athlete, nothing in the next and a loss in a third? The Unified Model of Tone answers with a single test it applies to every input, from a capsule to a surgery. It sets them on one axis and asks a single question of each: does the magnitude of this input match what this system, right now, can use?
A supplement is an input with a magnitude. The state it meets is what the model calls tone, and it is the variable the trials on this page never measured before dosing. Nitrate met an oxygen cost pathway already optimized by training. Caffeine met an enzyme that clears it at three different speeds. Creatine met stores that were full in some athletes and half empty in others. The compound was constant. The room was not.
The prediction this page makes
The model predicts that the size of an individual response is readable before the first dose, from the starting state of the system that compound acts on. Muscle creatine content predicts the creatine response Harris 1992. CYP1A2 genotype and habitual intake predict the caffeine response Guest 2018. Serum ferritin predicts who moves on iron Sim 2019.
The two beetroot responders among eight elite runners were identifiable in advance on this account, rather than discovered afterward in the individual data Boorsma 2014. That is a testable statement, and it inverts how supplement trials are usually designed: measure the state first, then dose, then read the response against the state.
Underneath sits the larger claim. The model holds that an athlete's separate readouts are windows on one regulatory state rather than independent talents, so an input that moves one is more likely to move the others in whoever responds, with compensation deciding which windows register it. This is a claim about how a supplement response is organized rather than a claim about what any product or any treatment does.
Run a squad through a standardized loading trial and record four measures in the same athletes, before and after. Those measures are RMSSD, reaction time variability, stride interval variability, and time to return to baseline after a repeated sprint test. If RMSSD, reaction time variability, stride interval variability and time to return to baseline move together in the athletes who respond, the unification claim is confirmed.
09The tone reading
What the Supplement Shelf Says About One Regulated System
Three foundations of tone carry the signature in the supplement literature, and each one already has a number attached to it.
Input quality
Purity is literally input quality. Of 634 products screened across 13 countries, 94 carried anabolic steroids that no label declared.
Set point
Creatine fills toward a ceiling. Uptake was largest in the muscles that started lowest, and in some of those subjects it rose by half.
Time course
Every compound here runs on a clock. Beta-alanine needs four weeks, bicarbonate 60 to 180 minutes, caffeine about an hour before the effort.
The remaining foundations show up in the same material. Gain is what caffeine changes, raising how strongly the system answers a given effort rather than how much force the muscle holds. Load is what a supplement is added on top of, and no capsule substitutes for the fuel and fluid the training itself demands. Constraint is what the buffers loosen, widening the range an athlete can work inside before acidity forces the pace down. Coupling is the nitrate story, where nitric oxide ties blood flow to the oxygen cost of the same work, so economy moves before finish time does. Prediction is the claim being made here, that a response is readable from the state measured before the first dose. Oscillation runs underneath creatine, which works on the fastest energy rhythm in the body, refilling ATP in the seconds between maximal efforts. The axis these compounds sit on, and the framework that holds it, is set out in the Unified Model of Tone.
10Where this sits
How This Page Relates to the Rest of the Library
Seven places this argument continues, each with the claim that earns the link.
Carries the carbohydrate and protein targets and the nutrient timing evidence this page deliberately does not restate.
Owns sodium, fluid balance and hyponatremia, and holds the hyperhydration protocols that glycerol belongs to.
Where low energy availability is handled, because no compound on this page compensates for eating too little.
The phosphocreatine, glycolytic and oxidative systems these compounds act on, taught as one continuous supply chain.
Holds the antioxidant literature, including the trials where high doses around training blunted the adaptation.
Owns RMSSD, the variability readout this page uses to state its prediction in measurable terms.
Treats anti-doping compliance as a scope and documentation question, which is where a supplement decision is recorded.
11Questions athletes ask
Questions Athletes Ask
Which supplements actually have evidence behind them for athletic performance?
Six compounds sit in the Australian Institute of Sport Group A performance column: creatine monohydrate, caffeine, beta-alanine, dietary nitrate as beetroot juice, sodium bicarbonate and glycerol. The IOC consensus statement named caffeine, creatine, specific buffering agents and nitrate as the few with good evidence behind them. Protein powder is classed as a sports food, which is convenient food rather than a true ergogenic aid. Total calories, macronutrient balance, micronutrient coverage and meal timing outrank every item on that list.
Which supplements help recovery between hard training sessions?
Creatine has the best case, because refilling the phosphocreatine pool supports repeated high intensity work and the training adaptations that follow. Protein supports repair, and its daily target belongs with sports nutrition rather than with ergogenic aids. The larger recovery variable is autonomic rather than chemical. The central integrative state an athlete carries into the next session decides how much parasympathetic recovery is available, and heart rate variability is how that gets measured. A buffer cannot fix a nervous system stuck in sympathetic overdrive.
How do I keep a supplement from triggering a positive doping test?
Buy only products that carry third-party batch testing, and check the batch rather than the brand. The reason is the contamination record: 14.8 percent of 634 supplements analyzed in one international study held undeclared anabolic steroids, and 38 percent of 66 high-risk products screened two decades later held undeclared doping substances. Article 2.1.1 of the World Anti-Doping Code makes the athlete responsible for whatever is in their sample, with no need to show intent. USADA states that no supplement is guaranteed risk free.
Why does the same supplement work for one athlete and not another?
Because the compound is only half of the experiment and the starting state is the other half. Creatine uptake was largest in the subjects whose muscle stores started lowest. Caffeine at 4 milligrams per kilogram made AA carriers of CYP1A2 6.8 percent faster over 10 kilometers and CC carriers 13.7 percent slower. Dietary nitrate produced a clear pooled effect in recreationally active men and no significant effect in athletes above 65 mL/kg/min. Same input, different room to move.
Is creatine safe to take long term?
The evidence on the category is strong. The position stand on creatine reports that short and long term supplementation was safe and well tolerated in healthy people, at intakes up to 30 grams a day for five years. The same review covers patient groups from infants to the elderly. That is a statement about the compound, not about a particular bottle, since contamination is a separate question answered by batch testing. What dose suits one athlete is a clinical question that belongs with their own bloodwork and history.
Does protein powder count as a supplement?
It is classified as a sports food rather than as a performance supplement, and that label describes what it does. Whey delivers leucine, the amino acid that triggers the mTOR pathway, in a form that travels in a bag. That makes it a delivery method for food an athlete could also eat, and the daily target and its distribution across servings sit with sports nutrition. Powder is convenience when appetite or schedule fails, and it is not an ergogenic aid.
How do supplements fit the Unified Model of Tone?
The model treats a supplement as an input with a magnitude, and it asks one question of every input: does this magnitude match what the system can use right now? That reframes the scattered trial record as information rather than noise, because a null result marks a system with no room rather than a compound with no action. It also makes a testable claim, that the size of a response is readable from the state measured before the first dose.
12The sources
References
14 primary sources, each linked to its record. Figures quoted on this page were checked against the published abstract.
Related evidence