The Personal Pharmacology of Sports Nutrition: Why Caffeine, Beetroot, and Creatine Work for Some Athletes and Not Others
Two cyclists train together five days a week. They’re within a watt of each other on the same bike, with similar weights, similar diets, and similar sleep habits. On race morning, they take the same 200mg caffeine pill sixty minutes before the start of a 10-kilometer time trial they’ve done together a dozen times. One of them shaves forty seconds off of her personal best. The other finishes twelve seconds slower.
Why did caffeine positively affect one of the cyclists while negatively affecting the other? The trial that produced the most influential data on this question — Guest and colleagues' 2018 study at the University of Toronto — gave proportional caffeine doses to 101 male athletes who happened to differ at a single letter of their DNA at a single location in a single gene. The athletes carrying two A's at that genetic location got faster by an average of about seven percent at the higher dose. The athletes carrying two C's got slower by an average of nearly fourteen percent. The "average effect" of caffeine in that study was a three percent improvement, the kind of number that ends up in meta-analyses, and the kind of number that hides the interesting mechanisms about what’s going on inside any individual person who takes caffeine before a race [1].
This article is not a story about caffeine. Caffeine is one of three layers we will work through in this article, albeit the most studied one. Complementing this genetic layer is a second layer of individualization found in your mouth — specifically, a community of bacteria that lives on the back of your tongue and decides, on your behalf, whether the nitrate in a glass of beetroot juice ever becomes the nitric oxide that opens up your blood vessels. Underneath that is a third layer farther down in your gastrointestinal tract: gut microbes that, in elite endurance athletes, appear to convert the lactate your muscles produce during hard efforts into a short-chain fatty acid that may act as a performance-enhancing molecule.
The point of this article is not that you should run a saliva-DNA test and a stool-microbiome test before your next race. Most of these tests, as we will see, are not as useful as their marketing suggests. The point is that the supplements you and your training partners may argue about — caffeine, beetroot juice, creatine, beta-alanine, sodium bicarbonate, vitamin D, iron, etc. — work through molecular mechanisms whose efficacy depends on three things at once: your personal genome, your oral microbiome, and your gut microbiome. Once you understand how these three layers interact, one person’s experience of "this supplement transformed my training" coexisting with another person’s experience of "I tried it and felt nothing" stops being a paradox. It becomes the expected outcome of pouring the same molecule into different molecular machines.
This article walks through what is actually known about that machinery, supplement by supplement, and ends with a practical framework for figuring out what works for you specifically.
Why Caffeine Doesn't Work the Same for Everyone
Caffeine is the most widely consumed psychoactive drug on earth, and one of only a handful of dietary supplements that the International Society of Sports Nutrition explicitly endorses as ergogenic across a broad range of exercise tasks [2]. Its mechanism of action in the nervous system is a textbook case of receptor antagonism. Adenosine is a small molecule that accumulates in the brain over the course of the day and binds to adenosine A1 and A2A receptors, where it acts as a brake on neural activity. Caffeine, which has a similar shape to adenosine, binds to the same receptors but does not activate them — it sits in the receptor's parking spot without turning the engine on. With the brake released, neurons fire more readily, perceived effort drops, and the rate of fatigue accumulation during exercise slows. This effect on perceived exertion is probably the largest single contributor to caffeine's ergogenic effect, and it is why caffeine helps in nearly every exercise modality that has been tested [2].
There are at least two other mechanisms that contribute, both relevant to muscle rather than brain. Caffeine binds to the ryanodine receptor (RyR1) in skeletal muscle and increases calcium release from the sarcoplasmic reticulum during contraction, which can amplify force output at a given motor neuron drive [3]. Caffeine also inhibits phosphodiesterase activity, raising intracellular cyclic AMP, which activates protein kinase A and increases lipolysis — the breakdown of adipose tissue into free fatty acids that can be used as fuel [4].
This neat textbook story breaks down the moment you give the same dose to two different people.
The most important source of inter-individual variation in caffeine response is a single nucleotide polymorphism in the CYP1A2 gene called rs762551, sometimes written as “-163A>C” [1, 5]. CYP1A2 encodes a liver enzyme that is responsible for roughly 95 percent of caffeine metabolism — it is the molecular machine that takes caffeine apart so it can be excreted. The A allele is associated with higher CYP1A2 activity (a "fast metabolizer"), and the C allele with lower activity (a "slow metabolizer"). About 40 percent of people are AA homozygotes, about 50 percent are heterozygotes (AC), and about 10 percent are CC homozygotes, although true frequencies vary substantially across populations.
The Guest et al. 2018 study mentioned in the opening was a randomized, double-blind, placebo-controlled crossover design in 101 male athletes who completed three 10-kilometer cycling time trials at 0, 2, and 4 mg/kg caffeine [1]. The headline result was the genotype-by-dose interaction. Athletes with the AA genotype got faster at both doses — 4.8 percent faster at 2 mg/kg and 6.8 percent faster at 4 mg/kg. Athletes with the AC genotype showed no significant effect at either dose. And athletes with the CC genotype — the slow metabolizers — were actually slower at the 4 mg/kg dose by 13.7 percent. This is not a small effect. A 13.7 percent slowdown over a 10K time trial is the difference between a typical performance and a performance you would talk about for years.
Subsequent studies have produced a more complicated picture. A 2023 systematic review with meta-analysis examining whether the CYP1A2 ergogenic effect generalizes found that, across a more heterogeneous body of literature, the genotype-stratified effect sizes are smaller than Guest's original numbers and depend heavily on exercise modality, dose, and habitual caffeine intake [6]. In some studies — particularly in resistance exercise rather than endurance time trials — the genotype interaction does not appear at all [7]. The most reasonable interpretation is that the CYP1A2 effect is real and largest in endurance contexts at moderate-to-high doses, but the magnitude observed in any individual study depends on factors the athlete cannot fully control.
A second variant that gets a lot of attention is rs5751876 in ADORA2A, the gene encoding the adenosine A2A receptor itself. The TT genotype at this position is associated with higher caffeine sensitivity, which sounds like it should predict a larger ergogenic effect. The empirical picture is messier. Some studies report that TT carriers experience greater performance benefit and stronger anti-inflammatory effects [8]. Others find that, paradoxically, C allele carriers — historically labeled "non-responders" — actually do show ergogenic responses to caffeine in well-designed trials, with magnitudes similar to non-genotyped populations [9]. Still others find no ADORA2A interaction at all, particularly in adolescents or in resistance exercise contexts [10]. The clinically clearer effect of ADORA2A variation is on the side-effect profile rather than the ergogenic effect: TT carriers are more likely to experience anxiety, jitters, and sleep disruption from a given caffeine dose. If your training partner can drink an espresso at 4 PM and sleep, and you cannot, the difference is probably here.
There are two practical takeaways. The first is that the standard ISSN caffeine recommendation — 3 to 6 mg/kg, ingested 60 minutes before exercise — is a population-level prescription [2]. Within that range, an AA genotype athlete probably benefits from the high end, and a CC genotype athlete probably benefits from the low end or from skipping caffeine entirely on race day. The second is that the only reliable way to figure out which group you are in, without genetic testing, is to test caffeine carefully under controlled conditions — same workout, same time of day, same nutrition, same warmup, with and without caffeine — and pay attention to performance rather than to subjective feel. Caffeine is a stimulant, and feeling more alert is not the same thing as being faster. If you have ever felt pumped up before a workout and then turned in an unimpressive set, you have already collected a data point on this distinction.
Beetroot Juice and the Bacteria on Your Tongue
Sometime around 2007, a small group at the Karolinska Institute in Stockholm published a paper that would eventually launch one of the most studied supplement protocols in endurance sport. Larsen and colleagues gave nine healthy men 0.1 millimoles of sodium nitrate per kilogram of body mass per day for three days — about the amount of nitrate you would find in 200 grams of spinach — and measured oxygen cost during submaximal cycling. The result was small, surprising, and counterintuitive: oxygen consumption at a given workload dropped by about five percent, with no change in lactate, suggesting the muscles had become more efficient at converting fuel into mechanical work [11]. A follow-up study from the same group in 2011, published in Cell Metabolism, traced the effect to the mitochondria themselves: dietary nitrate increased the P/O ratio — the amount of ATP produced per atom of oxygen consumed — in skeletal muscle mitochondria from human subjects [12].
This was the foundational work. Two years later, Bailey and colleagues at the University of Exeter took the next obvious step and asked whether the efficiency improvement translated into time trial performance in cyclists. Nine club-level male cyclists drank 500 ml of beetroot juice — about 6.2 millimoles of nitrate, roughly the amount you would get from a moderate serving of vegetables — about 2.5 hours before completing 4-kilometer and 16.1-kilometer cycling time trials. Both performances improved by about 2.7 to 2.8 percent, with the cyclists producing more power for the same oxygen cost [13]. The umbrella reviews now collectively show small but real ergogenic effects across high-intensity efforts of 5 to 30 minutes, with chronic supplementation generally outperforming acute dosing [14, 15].
The mechanism connecting "drink beetroot juice" to "improve cycling time trial" is one of the more peculiar pieces of human physiology, because it does not actually run through human enzymes. It runs through a community of bacteria that lives on your tongue.
The pathway is called the enterosalivary nitrate-nitrite-nitric oxide pathway, and it works like this: Dietary nitrate (NO₃⁻) — abundant in beetroot, spinach, arugula, and other leafy greens — is absorbed in the small intestine, enters the bloodstream, and is then actively concentrated in the salivary glands. About 25 percent of the nitrate you ingest ends up in your saliva, where the concentration can be ten times higher than in your plasma [16]. So far this is purely human physiology. The next step is not. Humans do not have a nitrate reductase enzyme — we cannot, on our own, convert nitrate (NO₃⁻) to nitrite (NO₂⁻). The bacteria living on the dorsal surface of our tongues can. Several genera — Veillonella, Rothia, Actinomyces, Neisseria, Haemophilus, and certain Streptococcus species — express bacterial nitrate reductase enzymes that reduce nitrate to nitrite as part of their own anaerobic respiration [17, 18]. The nitrite they produce is then swallowed back into the gut and absorbed.
Once in the bloodstream, nitrite is reduced one more step, to nitric oxide (NO), via several mammalian enzymes that act as nitrite reductases under the right conditions: xanthine oxidoreductase, deoxyhemoglobin in red blood cells, deoxymyoglobin in muscle, and aldehyde oxidase. Each of these enzymes reduces nitrite to NO most efficiently under low-oxygen, low-pH conditions — exactly the conditions inside a working muscle during high-intensity exercise [19]. The NO then activates soluble guanylate cyclase in vascular smooth muscle, raising cyclic GMP, relaxing the smooth muscle, and dilating the blood vessels supplying the working tissue. NO also appears to act inside the mitochondrion itself, where it modulates the efficiency of the electron transport chain — the apparent source of the P/O ratio improvement Larsen documented [12].
This is where the personalization story gets real. The bacteria on your tongue are doing a step in your physiology that you cannot do yourself, and the composition of that bacterial community varies widely between individuals.
The cleanest demonstration of this dependency comes from a 2008 paper by Govoni and colleagues, in which seven healthy volunteers were given a sodium nitrate load and had their plasma nitrite measured under two conditions: with their normal oral microbiome intact, and after rinsing with chlorhexidine antibacterial mouthwash [20]. The mouthwash, which is widely used in dentistry to suppress oral bacteria, reduced the rise in plasma nitrite by about 90 percent. The bacteria on the tongue are a load-bearing part of the pathway. Remove them, and the ergogenic effect of nitrate goes with them. Subsequent studies have replicated this finding using everyday over-the-counter mouthwashes and shown corresponding effects on blood pressure: people who use antibacterial mouthwash regularly have higher blood pressure than people who do not, all else equal, and the post-exercise blood pressure drop is blunted by chlorhexidine rinsing [21, 22].
The implication for athletes is direct: if you are using beetroot juice or another nitrate source as an ergogenic aid, you should not be using antibacterial mouthwash in the hours before exercise. The more interesting implication is that the people for whom beetroot juice does not work — and there are real non-responders, with estimates ranging from 25 to 40 percent in the literature — may not be failing to respond because of their genes. They may be failing to respond because the bacteria on their tongue do not include enough nitrate-reducing species, or because those species are present but suppressed by dental hygiene products, or because their oral pH is unfavorable for the bacterial reductase reaction.
This is where the science gets interesting, and where it is starting to point toward genuine personalization. Multiple studies have now shown that highly trained athletes have measurably different oral microbiomes from sedentary controls, with enrichment of the nitrate-reducing genera that drive the ergogenic effect. A 2025 review of seven studies found that athletes consistently show higher relative abundances of Veillonella and Rothia in tongue swabs, and that swimming, rugby, and water polo training in particular appear to remodel the tongue microbiome toward this functional profile [18]. A separate 2025 study in Scientific Reports compared tongue dorsum microbiomes between competitive athletes and inactive controls and found significantly different beta-diversity, with athletes enriched for nitrate-reducing Rothia mucilaginosa and unclassified Gemella species [23]. And a 2025 trial showed that eight weeks of high-intensity interval training in previously sedentary men was sufficient to alter the tongue microbiome and increase salivary nitrite levels — the bacterial community is responsive to training [24].
It is not yet clear whether having a "good" nitrate-reducing tongue microbiome causes athletic adaptation or whether athletic adaptation produces it, and the bidirectional relationship is hard to disentangle. What is clear is that the bacteria are part of the machinery, and that the machinery is more responsive than people thought a decade ago. For the practical athlete, two things follow. First, do not use antibacterial mouthwash before nitrate-dependent exercise — and probably do not use it routinely if you are training hard, because there is now reasonable evidence that mouthwash impairs the post-exercise blood pressure response and possibly some aspects of recovery [25]. Second, dietary nitrate works better when consumed chronically rather than acutely, in part because chronic intake appears to selectively enrich the relevant bacterial species in your own mouth. In other words, you are training your microbiome along with your muscles [26].
The Marathoner Microbe
In 2015, a group of researchers led by Jonathan Scheiman at Harvard Medical School collected stool samples from 15 runners who completed the Boston Marathon and from 10 sedentary control subjects, before and after the race. The runners' stool microbiomes were sequenced and compared to the controls and to each other across time points. The finding that emerged from that data set, published in Nature Medicine in 2019, made a strong claim about what the gut microbiome of an elite endurance athlete actually does, and the years since have begun to slowly flesh it out [27].
The headline observation was that one bacterial genus, Veillonella, was substantially enriched in the runners' stool samples in the days following the marathon compared to baseline, while no equivalent enrichment occurred in the sedentary controls. Veillonella species are interesting because of what they eat. Most gut bacteria use carbohydrates as their primary carbon source. Veillonella uses lactate. The same lactate that your muscles produce during high-intensity exercise — the same lactate whose accumulation is associated with the burning sensation in working muscles — turns out to be a metabolic substrate for one of the bacteria that thrives in marathon runners' guts.
Scheiman and colleagues isolated a strain of Veillonella atypica from one of the runners and gavaged it into mice. Compared to control mice that received Lactobacillus bulgaricus — the bacterium found in yogurt — the V. atypica-treated mice ran approximately 13 percent longer on a treadmill before exhaustion. The performance effect was real, robust, and reproducible. The researchers then asked: how is the bacterium making the mice faster?
The answer involves a metabolic pathway that bridges the gut and the rest of the body in a way that was not previously appreciated. V. atypica metabolizes lactate through the methylmalonyl-CoA pathway and produces propionate, a short-chain fatty acid, as the primary end product. Propionate is absorbed across the intestinal epithelium into systemic circulation, where it has a number of physiological effects — including, apparently, a performance-enhancing one. The Scheiman group tested the propionate hypothesis directly by administering propionate intracolonically to mice — bypassing the bacteria entirely — and found that propionate alone reproduced the increased treadmill time (p = 0.03). The mechanism is, at minimum, "lactate from working muscle reaches the gut, V. atypica converts it to propionate, propionate enters circulation, and propionate enhances performance." The remaining open questions, which the authors honestly acknowledged, are why Veillonella in particular is the lactate-utilizing genus that gets enriched (rather than other lactate-metabolizing organisms that exist), and exactly how circulating propionate enhances exercise performance — whether through energy substrate provision, anti-inflammatory effects, or some other mechanism [27].
The Scheiman paper has become foundational, but it has not been followed by a wave of "swallow some Veillonella and run faster" interventions, for reasons worth understanding. The first reason is that bacterial colonization of the gut is highly dependent on the existing microbial community and ecological niche. Introducing a strain that does not have a place to live is unlikely to produce stable colonization. The second is that Veillonella enrichment in marathon runners is probably an effect of training as much as a cause of performance, and the directionality of the relationship is hard to establish in human studies. The third is that "more Veillonella" is unlikely to be the universal answer; multiple lactate-utilizing genera exist, and the right community composition may depend on diet, training stimulus, and other factors. A 2025 narrative review of athletic gut microbiomes concluded that the relevant signal is not any single genus but a broader functional profile — short-chain fatty acid producers, lactate utilizers, and carbohydrate fermenters all enriched together — that appears to support endurance, recovery, and immune function [28].
What this means for an athlete is more modest than the press coverage of the original paper suggested. There is no commercially available probiotic that has been shown to replicate Scheiman's mouse result in humans. Marketers happy to sell "marathoner microbe" products have outpaced the science by about a decade. What does seem to be replicated, across multiple studies, is that high training loads consistently shift gut microbiome composition toward higher SCFA production, that this shift correlates with measures of endurance capacity, and that the shift is reversible — detraining returns the microbiome closer to a sedentary profile within weeks [28, 29]. The microbiome is part of the adaptation to training, not just a fixed trait you happen to have. As of 2026, the practical implication is mostly negative: avoid unnecessary courses of broad-spectrum antibiotics during training blocks, eat enough fiber to feed the SCFA producers, and don't expect the microbiome story to give you a shortcut around training. The slower, more honest implication is that the gut is going to turn out to be a third axis of personalization in sports performance, alongside the genome and the oral microbiome, and the mechanisms are starting to come into focus.
Other Supplements Through This Lens
The pharmacogenomic and microbiome stories that animate the caffeine and nitrate cases are the cleanest examples of personalization in sports nutrition, but they are not the only ones. The supplements that’ve made it into the small evidence-supported category — creatine, beta-alanine, sodium bicarbonate, vitamin D, and iron — each have their own story about why some athletes benefit substantially and others don't. In some cases the variance is genetic. In some cases it’s dietary. In some cases it’s structural, having to do with how the supplement is delivered to the body. And in at least one case, it’s honestly very small: not all supplements have meaningful inter-individual variation, and saying so is part of getting the personalization story right.
Creatine
Creatine is the supplement with the strongest aggregate evidence base in sports performance, and the supplement whose responder/non-responder distinction has been understood the longest. The 1994 study by Greenhaff and colleagues at the University of Nottingham — which used muscle biopsies from the vastus lateralis to measure intramuscular creatine content before and after five days of supplementation at 20 g/day — found that five of eight subjects increased their muscle total creatine substantially (15 to 32 percent) while three subjects showed minimal increases (5 to 7 percent) and no improvement in phosphocreatine resynthesis [30]. Subsequent work converged on a working definition of "non-responder" as a subject with less than 10 mmol/kg dry weight increase in muscle creatine after a standard loading protocol, with prevalence estimates of 20 to 30 percent in the population [31].
The mechanism is structural, not genetic in the SNP-variant sense. Skeletal muscle has a ceiling on intramuscular creatine, somewhere around 150 to 160 mmol per kilogram of dry muscle. The transporter responsible for moving creatine into muscle cells, encoded by the SLC6A8 gene, downregulates as intracellular creatine concentration approaches saturation [32]. People who walk into the supplement store with high baseline muscle creatine — typically meat-eaters whose diet provides 1 to 2 grams of creatine per day — have less room to fill up, and their measured response to a loading protocol is correspondingly small. People with low baselines — vegetarians and vegans, who get essentially no dietary creatine because plants do not synthesize it — start with intramuscular creatine concentrations 10 to 30 percent below omnivore baselines, have substantially more room to fill, and consequently show larger absolute and proportional gains from supplementation [33, 34].
The practical implication is the opposite of "test your genome to find out if you'll respond to creatine." The single best predictor of how much you will benefit from creatine supplementation is how much creatine you currently get from food. If you eat meat, fish, and dairy regularly, your baseline is probably high and your response will be moderate. If you have been a vegetarian for years, your baseline is probably low and your response will be substantial. The standard protocol — 20 grams per day in four 5-gram doses for five to seven days, followed by 3 to 5 grams per day chronically — works reliably across this spectrum, with effect size proportional to baseline depletion. There is no genetic test that adds meaningful information to this picture for healthy athletes, although there are rare SLC6A8 mutations that cause creatine transporter deficiency syndromes and a different clinical picture entirely.
Beta-Alanine
Beta-alanine is a useful supplement to talk about precisely because it does not show much pharmacogenomic variation, and the temptation to assume every supplement must have a genetic personalization story leads to claims that the data does not support. The mechanism is well understood: beta-alanine is the rate-limiting amino acid precursor to muscle carnosine, a dipeptide that buffers protons in working muscle. During high-intensity exercise, ATP regeneration through glycolysis produces a stoichiometric increase in intracellular H⁺, and intramuscular carnosine acts as one of the major proton buffers preventing the resulting drop in pH. Increasing muscle carnosine concentration via beta-alanine supplementation extends the duration of high-intensity exercise that can be sustained before pH-mediated fatigue, particularly for efforts in the 1- to 4-minute range [35, 36].
The dose-response curve is well characterized. Daily supplementation with 4 to 6 grams of beta-alanine for two to four weeks increases muscle carnosine by 20 to 30 percent in the first two weeks, 40 to 60 percent by four weeks, and approaches a plateau around 80 percent above baseline at ten weeks [37]. The 2015 ISSN position stand consolidates the evidence and recommends the same range [35].
Several genes encode enzymes involved in beta-alanine and carnosine metabolism — most notably CNDP1, which encodes serum carnosinase (the enzyme that breaks down carnosine in plasma), and CARNS1, which encodes carnosine synthase in muscle. It would be reasonable to hypothesize that polymorphisms in these genes modulate individual response to beta-alanine supplementation, and the supplement industry has been happy to suggest exactly this. The empirical reality is more deflating. The most thorough review of the question concluded that vegetarianism, female sex, and increasing age are associated with reduced muscle carnosine levels at baseline — but CNDP1 genotype is not a major determinant of muscle carnosine response to supplementation [38]. Most of the inter-individual variance in beta-alanine response is explained by training status, baseline muscle fiber composition, and dietary intake of beta-alanine-containing foods, not by genetic variants.
The practical takeaway is that beta-alanine is one of the few supplements where you can largely just follow the protocol — 4 to 6 grams per day, divided across smaller doses to manage paresthesia (the pins-and-needles sensation that is harmless but uncomfortable), for at least two to four weeks before expecting performance effects — and most people will respond. Vegetarians and women may respond more than average. People who eat large amounts of chicken, beef, or fish (which contain meaningful amounts of carnosine and its precursors) may respond less. Genetic testing does not currently add useful signal here, even though several DTC companies will sell you a beta-alanine "responder" panel.
Sodium Bicarbonate
Sodium bicarbonate is interesting because the inter-individual variation in response is real but is not primarily about genetics or microbiology. It is about plumbing — specifically, what happens when you ingest a substantial alkali load and your stomach reacts to it. The mechanism of action is straightforward: ingestion of 0.2 to 0.5 grams per kilogram of sodium bicarbonate, 60 to 90 minutes before exercise, raises extracellular bicarbonate concentration in the blood. The increased extracellular buffer capacity steepens the gradient for proton (H⁺) extrusion from working muscle, allowing higher intracellular pH for a given metabolic load. The 2021 ISSN position stand summarizes the performance evidence: sodium bicarbonate reliably improves performance in high-intensity exercise tasks lasting 30 seconds to 12 minutes, with the strongest effects in repeated-sprint, combat sport, and middle-distance contexts [39].
The historical problem with sodium bicarbonate has nothing to do with whether it works at the muscle level. It has to do with what happens in the stomach when you swallow 30-something grams of sodium bicarbonate dissolved in water. The standard 0.3 g/kg dose causes gastrointestinal distress in a substantial fraction of users — bloating, cramping, nausea, and the kind of urgency that ruins a warm-up. Athletes who tolerate the dose can extract a 1 to 2 percent performance benefit. Athletes who do not tolerate the dose either skip the supplement entirely or end up using a smaller, less effective dose. For decades, this delivery problem put sodium bicarbonate in a frustrating category: a supplement with strong mechanistic support and reasonable evidence of efficacy that many athletes simply could not use.
The Maurten Bicarb hydrogel system, introduced in the early 2020s and validated in several controlled trials starting in 2024, addresses this delivery problem directly. The system consists of small bicarbonate mini-tablets (3 mm diameter) ingested with a carbohydrate hydrogel that encapsulates the tablets and reduces their interaction with stomach acid. In a 2024 trial published in Sports Medicine, the hydrogel system reduced GI symptoms by approximately 80 percent compared to traditional capsule delivery while producing equivalent acid-base effects, and improved 4-kilometer cycling time trial performance by about 4 to 5 seconds [40]. A separate 2024 study reported a 1.42 percent improvement in 40-kilometer cycling time trial performance with the hydrogel system, again with minimal GI symptoms [41]. The personalization layer here is structural rather than genetic. People whose stomachs tolerate alkali loads can use the cheap traditional protocol. People whose stomachs do not now have a delivery technology that works for them.
Vitamin D
Vitamin D is not really an ergogenic aid in the sense that caffeine or nitrate are. Its role in athletic performance is permissive rather than active — being deficient is a measurable handicap, but being sufficient does not push you above baseline. What makes it relevant to a personalization discussion is the prevalence of deficiency in athletes and the genetic variation in vitamin D receptor (*VDR*) signaling.
Roughly 60 percent of elite athletes have vitamin D insufficiency (25-OH-D below 30 ng/mL) at some point in the year, and roughly 30 percent are frankly deficient (below 20 ng/mL) [42, 43]. The risk factors are unsurprising: latitude above 40°N, winter season, indoor sport (basketball, gymnastics, swimming, ice sports), darker skin pigmentation, and high training loads. Vitamin D plays roles beyond bone metabolism that are directly relevant to athletic performance — most notably in skeletal muscle, where vitamin D receptors are expressed in muscle fibers and where deficiency is associated with type II fiber atrophy and reduced strength.
Several common polymorphisms in VDR — typically referred to as FokI, BsmI, ApaI, and TaqI — modulate receptor function and have been studied in athletic populations. A 2025 study in elite Kazakhstani athletes reported that the G/G genotype of the TaqI polymorphism was associated with both vitamin D insufficiency and increased injury risk [44]. Other studies have found associations between VDR variants and competitive performance in track and field [45]. The effect sizes are modest, and the findings are not always replicated across populations.
The practical implication for an athlete is the inverse of the genetic-test pitch. The single most useful test in this domain is a blood panel measuring 25-hydroxyvitamin D, which tells you exactly where you are right now. A genetic test telling you that you have a VDR variant associated with lower vitamin D efficacy adds essentially nothing to the practical decision, which is "supplement until your blood level is in the 30 to 50 ng/mL range." Indoor athletes, athletes training at high latitudes during winter, and athletes with low dietary intake of fatty fish should test their levels regardless of genotype, and supplement based on the result.
Iron
Iron deficiency is one of the most common nutritional issues in endurance sport, particularly among female athletes. Prevalence estimates range from 15 to 35 percent in female endurance athletes generally, climbing to over 50 percent in some elite cohorts [46]. Iron is essential for oxygen transport (in hemoglobin), oxygen storage in muscle (in myoglobin), and mitochondrial energy production (in iron-sulfur clusters of the electron transport chain). Iron deficiency, with or without anemia, measurably reduces endurance performance and recovery capacity.
The molecular regulator that ties iron status to athletic performance most tightly is hepcidin, a peptide hormone produced by the liver that regulates iron absorption from the gut and iron release from storage. Hepcidin levels rise in response to inflammation (mediated by IL-6) and to high body iron stores, and they fall when the body needs more iron. The relevance to athletes is that intense exercise — particularly weight-bearing endurance exercise — produces a transient inflammatory response that elevates hepcidin for several hours afterward. Studies show that post-exercise hepcidin surges peak around 6 hours after exercise and reduce dietary iron absorption by 30 to 50 percent during that window [47]. Athletes who routinely take iron supplements with their post-workout meal are getting substantially less iron than the dose suggests. Iron should be taken either well before exercise or 4 or more hours after, ideally with vitamin C to enhance absorption, and ideally not with coffee, tea, or calcium-rich foods, which inhibit absorption.
The genetic angle here involves HFE, the gene whose mutations cause hereditary hemochromatosis. The C282Y and H63D variants in HFE affect iron sensing and storage, with homozygotes for the most severe variant at risk of iron overload over decades. Because iron deficiency is the more common athletic problem, HFE genetic testing is rarely the right first step. The right first step is a blood panel — ferritin, transferrin saturation, complete blood count — that tells you your actual iron status. Ferritin levels below 30 ng/mL indicate depleted stores even when hemoglobin is in the normal range, and in symptomatic athletes (unexplained fatigue, declining performance, frequent illness) ferritin levels below 50 ng/mL may warrant intervention [46, 47].
What You Can Actually Test — and What's Hype
The DTC genetic testing industry has grown substantially over the past decade, and a large fraction of its marketing focuses on athletic performance and personalized nutrition. Companies will sell you a saliva-DNA test that promises to tell you whether you should eat more carbs or fats, whether you'll respond well to endurance or power training, whether caffeine will help or harm your performance, and so on. Most of these claims are not supported by the underlying evidence.
The most thorough evaluation of this question came in a 2015 consensus statement published in the British Journal of Sports Medicine and authored by a panel including Webborn, Williams, McNamee, and others [48]. The panel surveyed the DTC sports genetic testing market and the underlying evidence and concluded, in their published language, that "genetic tests have no role to play in talent identification or the individualised prescription of training to maximise performance." They were specific about the most-tested variant: ACTN3 R577X, the so-called "sprinter gene," is tested by 89 percent of DTC sports genetic testing companies despite the fact that the polymorphism explains less than 1 percent of the variance in sprint performance in the general population. The panel recommended that "no child or young athlete should be exposed to DTC genetic testing to define or alter training or for talent identification," based on a combination of poor predictive validity, ethical concerns, and quality control failures (different DTC companies returning different results from identical samples) [48].
A 2017 review by Pickering and Kiely argued for a more nuanced position on ACTN3 specifically — the gene appears to influence not just sprint performance but also muscle damage from eccentric training, recovery, and injury risk, with the R allele associated with enhanced strength gains and protection from training-induced damage [49]. But even this more sympathetic reading of the ACTN3 literature comes with the caveat that effect sizes are small, study populations are heterogeneous, and the practical implications for an individual athlete's training prescription are minimal. ACTN3 status is not actionable in the way that, say, your post-workout fatigue or your training tolerance is actionable.
Within the supplement domain specifically, the variants with the strongest evidence — the ones where genotype meaningfully changes the practical recommendation — are a short list:
CYP1A2 rs762551 for caffeine, where AA homozygotes consistently benefit from doses in the 4 to 6 mg/kg range and CC homozygotes may want to limit themselves to lower doses or skip caffeine on race day [1, 6].
DR polymorphisms for vitamin D, where the effect is modest and a blood test almost always provides better information than the genotype alone [44, 45].
HFE variants for iron metabolism, where the relevant clinical question is iron overload risk over decades rather than acute athletic performance, and where ferritin and transferrin saturation panels are the actionable measurements [46].
What this suggests is a hierarchy. For most supplements and nutrients, a $50 to $100 blood panel — measuring ferritin, transferrin saturation, complete blood count, 25-hydroxyvitamin D, and possibly thyroid and metabolic panels — provides substantially more useful, actionable information than a $200 genetic test. The blood panel measures your current state. The genetic test estimates a stable predisposition that may or may not be expressed given your current diet, training, and environment. For caffeine specifically, a one-time genetic test can be informative, although structured self-testing — measured time trials with and without caffeine, controlled for other variables — yields the same practical answer at zero genetic-counseling risk.
The microbiome testing market is even less mature. Stool microbiome tests are widely available, and several companies will sell you a "performance microbiome" analysis that scores your Veillonella abundance, your SCFA producer profile, or your "endurance bacteria" enrichment. The science underlying these scores is mostly inferred from cross-sectional studies in small athlete populations, with no validated longitudinal evidence that intervening on the score in a particular individual will produce a meaningful performance change. The 2019 Scheiman result is real and has held up, but it has not generated a validated commercial product. As of 2026, microbiome testing in athletes is best understood as a research tool, not a personalization tool.
The honest summary for most athletes is that the cheapest and most useful piece of personalized information you can buy is a comprehensive blood panel, particularly if you have any symptoms suggesting deficiency — unexplained fatigue, declining performance, frequent illness, slow recovery, mood changes. The second most useful piece of information is structured self-testing of supplements that the literature suggests should work: take it for two weeks, stop for two weeks, take it again, and pay attention to objective metrics. A third tier is selective genetic testing — CYP1A2 for caffeine is the clearest case — but only after the blood panel basics are dialed in.
A Practical Framework: Testing on Yourself
The pharmacogenomics literature, taken seriously, leads to a counterintuitive conclusion. Even when a genotype-supplement interaction is real and replicated — as it is for CYP1A2 and caffeine — the within-genotype variation across individuals is often as large as the between-genotype effect. Two AA homozygotes can have substantially different caffeine responses because they differ in habituation, sleep status, training context, and a dozen other factors that no genetic test captures. The published "AA genotype responds 6.8 percent" effect is the average within a heterogeneous group, not a property of any specific individual.
The implication is that even a perfect genetic test would not eliminate the need for self-testing. It would only suggest a starting point. The reliable way to figure out what works for you specifically is to run controlled experiments on yourself, with the same rigor that you would expect from a published trial.
The structure of an N-of-1 supplement test is straightforward in principle and harder to execute in practice. The principle: pick a single supplement, define a single performance test, and run the test multiple times with and without the supplement under conditions that are as identical as possible. The harder part: making the conditions actually identical. Sleep affects performance more than most supplements. Food affects performance more than most supplements. Stress, hydration, training accumulation, ambient temperature, and your psychological state all contribute variance that, in an uncontrolled comparison, can swamp the effect you are trying to measure.
Some practical guidelines that emerge from the methodological literature on athlete self-testing:
Pick an unambiguous performance metric. Time over a fixed distance, average power over a fixed duration, total volume at a fixed RPE, reps to failure at a fixed load. Avoid subjective metrics like "feel" or "energy," which are dominated by the placebo effect and by your expectations. If your best measurement is a vague sense of how the workout went, you cannot reliably detect supplement effects smaller than about 10 percent — and most real supplement effects are 1 to 5 percent.
Run multiple trials in each condition. A single comparison of "I tried it once and felt great" is an anecdote, not evidence. You want at least three to five trials in each condition, alternating in a structured way (ABBA or ABAB rather than AAA-BBB, to control for fitness drift across the testing period).
Control the obvious confounders. Same time of day. Same warm-up protocol. Same nutrition in the 24 hours before. Same sleep duration the night before, to within an hour. Same hydration. Same caffeine intake (if you are not testing caffeine). Same shoes, same equipment, same route or test. Anything you do not control is contributing noise that obscures the effect.
Blind yourself if possible. This is harder for supplements with obvious sensory effects — caffeine, sodium bicarbonate — but possible for tasteless or capsule-form supplements. A friend or coach can prepare matched placebo and treatment doses without telling you which is which. The placebo effect on self-reported energy is large; the placebo effect on objective performance is real but usually smaller.
Log it. Keep a written record. Memory is unreliable, and the brain selectively remembers the trials where the supplement seemed to work.
Recognize when to stop. If, after a structured comparison with adequate trials, you cannot detect a difference, the supplement probably is not doing what its marketing promises in your particular case. The right response is to stop spending money on it. The wrong response is to keep taking it because the evidence is "mixed" — the evidence in your case is what your testing produced.
Two cautions about self-testing matter. The first is that there are supplements where the effect is dose-cumulative — beta-alanine, creatine, vitamin D — and a two-week comparison will not detect a real effect that requires four to twelve weeks to develop. For these supplements, the testing protocol has to be longer, and the controls are correspondingly harder to maintain. The second is that some performance-relevant nutrients — vitamin D, iron, ferritin — are best assessed by blood tests rather than performance tests. If your ferritin is 18 ng/mL, you do not need to test whether iron supplementation improves your tempo runs; you need to address a deficiency.
When to stop tinkering altogether: if your training, sleep, nutrition basics, and stress management are not dialed in, supplements are noise on top of variance you have not yet controlled. The reason most athletes do not see consistent supplement effects is not that the supplements are useless — although some are — but that the day-to-day variance in their training, recovery, and life is large enough to swamp any supplement signal smaller than about 5 percent. Almost all supplements produce smaller-than-5-percent effects in well-trained athletes. Get the basics first; then run controlled tests on the supplements that, mechanistically, have a reason to work.
Honest Limitations
This article has presented the research on sports nutrition pharmacogenomics and microbiome science with what I hope is a fair degree of optimism — these fields are real, the mechanisms are getting clearer, and the practical implications are starting to be actionable. But several caveats are important for any reader who wants to use this material to make decisions about their own training.
Most pharmacogenomics studies in this space are small. Guest et al. 2018 had 101 athletes [1]. The original Greenhaff creatine non-responder study had 8 subjects [30]. The Scheiman V. atypica study sequenced 25 stool samples [27]. Single-study findings, even with strong effect sizes, do not become reliable until they are replicated by independent groups. The CYP1A2-caffeine literature has accumulated enough replication to be reasonably credible. The ADORA2A literature has not. The microbiome-performance literature is mostly cross-sectional and observational, with very few interventional trials. As you move from "well-established" to "interesting recent finding," the probability that the result is real, generalizable, and applicable to your specific case drops substantially.
Effect sizes are usually modest even when statistically significant. Most well-tested ergogenic supplements produce 1 to 5 percent improvements in performance under ideal conditions. The placebo effect in sports nutrition trials runs around 1 to 3 percent. The day-to-day variance in athletic performance from sleep, stress, and accumulated training is typically 3 to 8 percent. Distinguishing a 2 percent supplement effect from background noise requires careful experimental design, multiple trials, and honest accounting for confounders. If a supplement company is selling you a 20 percent improvement, the most likely explanation is selection bias in the testimonials.
Replication matters more than first findings. The literature is full of interesting first studies that failed to replicate. The DTC genetic testing market has been built on associations that look strong in single studies and dissolve when tested independently. The principle here is the same as in any other area of medical science: be skeptical of findings that have not been replicated by independent groups, particularly when those findings support the marketing claims of the people selling you the test.
There are real industry incentives to oversell personalization. Selling you a $200 genetic test that recommends supplements you might or might not benefit from is a more profitable business than selling you a $50 blood test that shows your iron is fine and you do not need to spend money on iron supplements. The personalization framing is intrinsically appealing — everyone wants to believe they are uniquely optimized — and the supplement industry has discovered that "personalized" sells better than "generic." Some of the resulting products are genuinely useful. Many are not.
Genotype-environment interactions are real and underdiscussed. The CYP1A2 effect on caffeine response is largest in athletes who do not habitually consume large amounts of caffeine. Habituation appears to attenuate or even eliminate the genotype effect for chronic high-dose users. A genetic test that reports you as an "AA fast metabolizer" without asking how much coffee you currently drink is delivering an incomplete picture. Similar interactions exist for vitamin D (where time outdoors and skin pigmentation matter), iron (where dietary intake and menstrual blood loss matter), and most other nutrient-related variants.
Most of the variance in athletic performance is not explained by any of this. Training, sleep, recovery, stress management, technical skill, and genetic variation in dozens to thousands of small-effect variants together explain orders of magnitude more performance variance than any single supplement-by-genotype interaction. The pharmacogenomics layer is real, but it is a small layer. The microbiome layer is real, but it is even smaller and less well understood. The big levers are still where they always were: how hard you train, how well you recover, how consistently you eat, and how much sleep you get.
Conclusion
The caffeine, beetroot, and creatine on a sports nutrition store shelf are the same molecules they have always been. What has changed — over the last fifteen years, more or less — is our understanding of what happens after those molecules cross the threshold of your body. Caffeine reaches a liver where one or another version of an enzyme breaks it down at one or another rate, and the difference between a four percent improvement and a fourteen percent decrement comes down to which version you happen to carry. Beetroot juice reaches a tongue where one or another community of bacteria converts the nitrate it contains to the nitrite that becomes nitric oxide, and the difference between responder and non-responder may depend less on your genome than on whether you have been using mouthwash. The lactate produced by your muscles during a hard interval reaches a gut where bacteria you did not choose convert it — or fail to convert it — into a short-chain fatty acid that may itself be a performance-enhancing molecule.
These three layers — pharmacogenomic, oral microbiome, gut microbiome — together describe a system that is far more variable across individuals than the published "average effect" of any supplement suggests. The fact that the standard caffeine recommendation works on average is a population-level claim. The fact that it works for you, at the dose you have been using, is a claim only your own controlled testing can establish. The science underlying personalization is real and is becoming more actionable each year. The commercial products that promise to operationalize that science are mostly running ahead of the evidence, and the most useful tool an athlete has — a blood panel, a structured self-test, a training log honest enough to capture what actually changed — has not changed in decades.
The frontier of sports nutrition is not a more sophisticated genetic test. It is the slow, unglamorous work of treating each athlete as a biochemically distinct individual, paying attention to what actually changes when something is tried, and discarding what does not work without ceremony. The molecular understanding now exists to explain why two cyclists who train together can take the same caffeine pill and produce opposite results. The job of using that understanding is, as it has always been, the athlete's own.
References
Guest N, Corey P, Vescovi J, El-Sohemy A. Caffeine, CYP1A2 Genotype, and Endurance Performance in Athletes. Med Sci Sports Exerc. 2018;50(8):1570-1578. doi:10.1249/MSS.0000000000001596
Guest NS, VanDusseldorp TA, Nelson MT, et al. International society of sports nutrition position stand: caffeine and exercise performance. J Int Soc Sports Nutr. 2021;18(1):1. doi:10.1186/s12970-020-00383-4
Murayama T, Kurebayashi N, Ogawa H, et al. Structural and functional interactions between the Ca2+-, ATP-, and caffeine-binding sites of skeletal muscle ryanodine receptor (RyR1). J Biol Chem. 2021;297(3):101040. doi:10.1016/j.jbc.2021.101040
Graham TE. Caffeine and exercise: metabolism, endurance and performance. Sports Med. 2001;31(11):785-807. doi:10.2165/00007256-200131110-00002
Sachse C, Brockmöller J, Bauer S, Roots I. Functional significance of a C-->A polymorphism in intron 1 of the cytochrome P450 CYP1A2 gene tested with caffeine. Br J Clin Pharmacol. 1999;47(4):445-449. doi:10.1046/j.1365-2125.1999.00898.x
Carvalho A, Marticorena FM, Grecco BH, Barreto G, Saunders B. Does ergogenic effect of caffeine supplementation depend on CYP1A2 genotypes? A systematic review with meta-analysis. J Sport Health Sci. 2024;13(2):220-231. doi:10.1016/j.jshs.2023.10.005
Klein CS, Clawson A, Martin M, et al. The effect of caffeine on cognitive performance is influenced by CYP1A2 but not ADORA2A genotype, yet neither genotype affects exercise performance in healthy adults. Eur J Appl Physiol. 2020;120(7):1525-1537. doi:10.1007/s00421-020-04384-8
Loy BD, O'Connor PJ, Lindheimer JB, Covert SF. ADORA2A C Allele Carriers Exhibit Ergogenic Responses to Caffeine Supplementation. Nutrients. 2020;12(3):741. doi:10.3390/nu12030741
Muñoz A, López-Samanes Á, Domínguez R, Moreno-Pérez V, Sánchez-Oliver AJ, Del Coso J. Effects of CYP1A2 and ADORA2A Genotypes on the Ergogenic Response to Caffeine in Professional Handball Players. Genes (Basel). 2020;11(8):933. doi:10.3390/genes11080933
Spineli H, Pinto MP, Dos Santos BP, et al. The effect of caffeine on exercise performance is not influenced by ADORA2A genotypes, alone or pooled with CYP1A2 genotypes, in adolescent athletes. Eur J Nutr. 2023;62(3):1041-1050. doi:10.1007/s00394-022-03051-2
Larsen FJ, Weitzberg E, Lundberg JO, Ekblom B. Effects of dietary nitrate on oxygen cost during exercise. Acta Physiol (Oxf). 2007;191(1):59-66. doi:10.1111/j.1748-1716.2007.01713.x
Larsen FJ, Schiffer TA, Borniquel S, et al. Dietary inorganic nitrate improves mitochondrial efficiency in humans. Cell Metab. 2011;13(2):149-159. doi:10.1016/j.cmet.2011.01.004
Lansley KE, Winyard PG, Bailey SJ, et al. Acute dietary nitrate supplementation improves cycling time trial performance. Med Sci Sports Exerc. 2011;43(6):1125-1131. doi:10.1249/MSS.0b013e31821597b4
Domínguez R, Cuenca E, Maté-Muñoz JL, et al. Effects of Beetroot Juice Supplementation on Cardiorespiratory Endurance in Athletes. A Systematic Review. Nutrients. 2017;9(1):43. doi:10.3390/nu9010043
López-Samanes Á, Pérez-López A, Morencos E, et al. Dietary Nitrate Supplementation and Exercise Performance: An Umbrella Review of 20 Published Systematic Reviews with Meta-analyses. Sports Med. 2025;55(8):1947-1968. doi:10.1007/s40279-025-02194-6
Lundberg JO, Weitzberg E, Gladwin MT. The nitrate-nitrite-nitric oxide pathway in physiology and therapeutics. Nat Rev Drug Discov. 2008;7(2):156-167. doi:10.1038/nrd2466
Hyde ER, Andrade F, Vaksman Z, et al. Metagenomic analysis of nitrate-reducing bacteria in the oral cavity: implications for nitric oxide homeostasis. PLoS One. 2014;9(3):e88645. doi:10.1371/journal.pone.0088645
Vidal-Vázquez NF, Hernández-Pérez R, Toledano-Ramos B, et al. Correlation Between the Oral Microbiota and Sports Practice: A Systematic Review. Cureus. 2025;17(1):e78168. doi:10.7759/cureus.78168
Lundberg JO, Weitzberg E. NO generation from nitrite and its role in vascular control. Arterioscler Thromb Vasc Biol. 2005;25(5):915-922. doi:10.1161/01.ATV.0000161048.72004.c2
Govoni M, Jansson EA, Weitzberg E, Lundberg JO. The increase in plasma nitrite after a dietary nitrate load is markedly attenuated by an antibacterial mouthwash. Nitric Oxide. 2008;19(4):333-337. doi:10.1016/j.niox.2008.08.003
Bondonno CP, Liu AH, Croft KD, et al. Antibacterial mouthwash blunts oral nitrate reduction and increases blood pressure in treated hypertensive men and women. Am J Hypertens. 2015;28(5):572-575. doi:10.1093/ajh/hpu192
Joshipura K, Muñoz-Torres F, Fernández-Santiago J, Patel RP, Lopez-Candales A. Over-the-counter mouthwash use, nitric oxide and hypertension risk. Blood Press. 2020;29(2):103-112. doi:10.1080/08037051.2019.1680270
Burleigh M, Liddle L, Monaghan C, et al. A comparison of oral microbiome composition between highly trained competitive athletes and untrained controls. Sci Rep. 2025;15(1):31845. doi:10.1038/s41598-025-16835-3
Burleigh M, Liddle L, Muggeridge DJ, et al. Eight weeks of high-intensity interval training alters the tongue microbiome and impacts nitrate and nitrite levels in previously sedentary men. Free Radic Biol Med. 2025;221:84-93. doi:10.1016/j.freeradbiomed.2025.02.005
Cutler C, Kiernan M, Willis JR, et al. Post-exercise hypotension and skeletal muscle oxygenation is regulated by nitrate-reducing activity of oral bacteria. Free Radic Biol Med. 2019;143:252-259. doi:10.1016/j.freeradbiomed.2019.07.035
Vanhatalo A, Blackwell JR, L'Heureux JE, et al. Nitrate-responsive oral microbiome modulates nitric oxide homeostasis and blood pressure in humans. Free Radic Biol Med. 2018;124:21-30. doi:10.1016/j.freeradbiomed.2018.05.078
Scheiman J, Luber JM, Chavkin TA, et al. Meta-omics analysis of elite athletes identifies a performance-enhancing microbe that functions via lactate metabolism. Nat Med. 2019;25(7):1104-1109. doi:10.1038/s41591-019-0485-4
Mancin L, Wu GD, Paoli A. Gut microbiota-bile acid-skeletal muscle axis. Trends Microbiol. 2023;31(3):254-269. doi:10.1016/j.tim.2022.10.003
Bongiovanni T, Yin MOL, Heaney L. The Athlete and Gut Microbiome: Short-chain Fatty Acids as Potential Ergogenic Aids for Exercise and Training. Int J Sports Med. 2021;42(13):1143-1158. doi:10.1055/a-1524-2095
Greenhaff PL, Bodin K, Soderlund K, Hultman E. Effect of oral creatine supplementation on skeletal muscle phosphocreatine resynthesis. Am J Physiol. 1994;266(5 Pt 1):E725-730. doi:10.1152/ajpendo.1994.266.5.E725
Syrotuik DG, Bell GJ. Acute creatine monohydrate supplementation: a descriptive physiological profile of responders vs. nonresponders. J Strength Cond Res. 2004;18(3):610-617. doi:10.1519/12392.1
Snow RJ, Murphy RM. Creatine and the creatine transporter: a review. Mol Cell Biochem. 2001;224(1-2):169-181. doi:10.1023/a:1011908606819
Kreider RB, Kalman DS, Antonio J, et al. International Society of Sports Nutrition position stand: safety and efficacy of creatine supplementation in exercise, sport, and medicine. J Int Soc Sports Nutr. 2017;14:18. doi:10.1186/s12970-017-0173-z
Kaviani M, Shaw K, Chilibeck PD. Benefits of Creatine Supplementation for Vegetarians Compared to Omnivorous Athletes: A Systematic Review. Int J Environ Res Public Health. 2020;17(9):3041. doi:10.3390/ijerph17093041
Trexler ET, Smith-Ryan AE, Stout JR, et al. International society of sports nutrition position stand: Beta-Alanine. J Int Soc Sports Nutr. 2015;12:30. doi:10.1186/s12970-015-0090-y
Saunders B, Elliott-Sale K, Artioli GG, et al. β-alanine supplementation to improve exercise capacity and performance: a systematic review and meta-analysis. Br J Sports Med. 2017;51(8):658-669. doi:10.1136/bjsports-2016-096396
Stellingwerff T, Anwander H, Egger A, et al. Effect of two β-alanine dosing protocols on muscle carnosine synthesis and washout. Amino Acids. 2012;42(6):2461-2472. doi:10.1007/s00726-011-1054-4
Rezende NS, Swinton P, de Oliveira LF, et al. The Muscle Carnosine Response to Beta-Alanine Supplementation: A Systematic Review With Bayesian Individual and Aggregate Data E-Max Model and Meta-Analysis. Front Physiol. 2020;11:913. doi:10.3389/fphys.2020.00913
Grgic J, Pedisic Z, Saunders B, et al. International Society of Sports Nutrition position stand: sodium bicarbonate and exercise performance. J Int Soc Sports Nutr. 2021;18(1):61. doi:10.1186/s12970-021-00458-w
Gough LA, Sparks SA, McNaughton LR, et al. The Effects of a Novel Sodium Bicarbonate Ingestion System on Repeated 4 km Cycling Time Trial Performance in Well-Trained Male Cyclists. Sports Med. 2024;54(11):2937-2949. doi:10.1007/s40279-024-02083-4
Oliveira LF, de Salles Painelli V, Nemezio K, et al. The effect of sodium bicarbonate mini-tablets ingested in a carbohydrate hydrogel system on 40 km cycling time trial performance and metabolism in trained male cyclists. J Int Soc Sports Nutr. 2024;21(1):2380057. doi:10.1080/15502783.2024.2380057
Owens DJ, Allison R, Close GL. Vitamin D and the Athlete: Current Perspectives and New Challenges. Sports Med. 2018;48(Suppl 1):3-16. doi:10.1007/s40279-017-0841-9
de la Puente Yagüe M, Collado Yurrita L, Ciudad Cabañas MJ, Cuadrado Cenzual MA. Role of Vitamin D in Athletes and Their Performance: Current Concepts and New Trends. Nutrients. 2020;12(2):579. doi:10.3390/nu12020579
Daulet B, Gusev O, Apsalikov B, et al. The association of VDR gene polymorphisms with serum vitamin D levels and injury predisposition in elite athletes of Kazakhstan. Eur J Appl Physiol. 2025;125(11):2895-2907. doi:10.1007/s00421-025-05976-y
Maciejewska-Skrendo A, Leźnicka K, Leońska-Duniec A, et al. Evaluation of the Association of VDR rs2228570 Polymorphism with Elite Track and Field Athletes' Competitive Performance. Genes (Basel). 2023;14(3):690. doi:10.3390/genes14030690
Sim M, Garvican-Lewis LA, Cox GR, et al. Iron considerations for the athlete: a narrative review. Eur J Appl Physiol. 2019;119(7):1463-1478. doi:10.1007/s00421-019-04157-y
Peeling P, Sim M, Badenhorst CE, et al. Iron Status and the Acute Post-Exercise Hepcidin Response in Athletes. PLoS One. 2014;9(3):e93002. doi:10.1371/journal.pone.0093002
Webborn N, Williams A, McNamee M, et al. Direct-to-consumer genetic testing for predicting sports performance and talent identification: Consensus statement. Br J Sports Med. 2015;49(23):1486-1491. doi:10.1136/bjsports-2015-095343
Pickering C, Kiely J. ACTN3: More than Just a Gene for Speed. Front Physiol. 2017;8:1080. doi:10.3389/fphys.2017.01080