Sodium
bicarbonate

It is the same compound as baking soda, NaHCO₃: the white powder that makes cakes rise. Chemically there is no separate sports version.

Your body already makes it and circulates it in the blood, where it is one of the central pieces of the system that keeps the body’s pH stable. Taking it in quantity raises blood bicarbonate for a few hours, and it is precisely in that rise that its whole effect lies.

It is not a stimulant and, with preserved kidney function, it does not accumulate: there is no need to take it daily. Nor do you feel anything when you swallow it, beyond the salty taste and, sometimes, its effects on the stomach.

The word that defines its function is buffer: a substance that absorbs hydrogen ions and cushions changes in acidity, just as a car’s shock absorber does not prevent the bump, but softens what reaches the cabin.

What sets it apart from other supplements with a similar purpose is where it works: outside the muscle cell, in the blood and in the fluid that surrounds it. It does not enter the fiber. It increases the acidity gradient between the inside and the outside, and that gradient pushes hydrogen ions outward.

Hence the effect lasts hours and not weeks: the kidney returns the excess to its usual value in a short time.

During a very intense effort, the muscle generates hydrogen ions faster than it can get rid of them, and that build-up is part of why you have to ease off the pace. The extra bicarbonate circulating in the blood helps draw those ions out of the muscle.

The improvement measured is small: meta-analyses place it between 2 and 3% in high-intensity exercise. In an event decided by tenths of a second, that counts; in any ordinary training session, you will not notice it.

The Australian Institute of Sport classifies it in Group A, its category of solid scientific evidence for specific sporting uses with established protocols.

About the improvement figure. The AIS’s 2-3% summarizes several meta-analyses. One of the most cited, by Carr and colleagues in 2011, gives 1.7%, with a 90% confidence limit of ±2.0%: an interval that includes zero and corresponds to a very specific condition — a single one-minute sprint, male athletes. They do not contradict each other: one is an institutional average and the other a point estimate with its uncertainty on display.

There is a second, less well-known pathway. The sodium that comes with the bicarbonate expands plasma volume, and it has been proposed that this hyperhydration contributes to performance by itself under some conditions (Siegler et al., 2021). It is a hypothesis with partial support, not an established mechanism.

And a third line, this one about training rather than competition: several works describe that supplementing during blocks of interval sessions amplifies aerobic adaptations — mitochondrial and oxidative function — beyond the acute effect on the day of the session (Edge et al., 2006; Percival et al., 2015). The AIS records it as growing evidence, not as a settled fact.

Its territory is high-intensity efforts lasting between 30 seconds and 10 minutes, such as the 100 and 200 meter swim, the 2,000 meter row, middle-distance running or track cycling. It can also be useful in combat, team and racket sports, where intense bursts are repeated with short rests.

Outside that window the backing blurs. In a ten-second sprint, the effort ends before acidity gets to be the problem; in a long race at a sustained pace, the limit is a different one: fuel and oxygen.

In an audit of the published studies on bicarbonate, only 4% had been done with women participants. The little that has been published points in the same direction, but it is too little to claim that the response is the same.

The window used by the position stand of the International Society of Sports Nutrition runs from 30 seconds to 12 minutes, with two documented extensions: events of around an hour raced just below threshold, where the extra buffering capacity can sustain a change of pace or a final sprint; and repeated intense efforts with short recoveries, typical of team, racket and combat sports.

The 2025 continuous running meta-analysis is often cited as proof that bicarbonate is no use for long endurance. But that is not what it analyzed: its tests lasted from 1 to 30 minutes, with a median of 4, in the heart of the good zone. What it calls into question is not the periphery, it is the center.

Strength and muscular endurance. The meta-analysis by Grgic et al. (2020) separates two things that are usually confused: bicarbonate improves muscular endurance — repetitions with a given load, total isokinetic work or time holding an isometric contraction — with a small effect size (SMD 0.37; 95% CI 0.15–0.59; p=0.001), but not maximal strength (SMD −0.03; 95% CI −0.18 to 0.12), neither at rest nor in a fatigued state. The authors themselves warn that task duration does not by itself explain where the effect appears: among the studies included there were tasks of more than 30 seconds with a null effect, and no meta-regression on duration was performed. Of 192 participants in total, 8 were women.

Varovic et al. (2023) is a randomized, double-blind trial with 19 participants, with a sodium chloride placebo at 0.21 g/kg — matched for sodium, which strengthens the result — on three sets at 70% of one-repetition maximum to failure in bench press and biceps curl. It is not an independent replication: two of its authors also sign the 2020 meta-analysis.

Competitions with repeated events. When there are heats and a final on the same day, the standard acute protocol does not fit. The strategies described are splitting the load or “topping up” with small doses at the start of recovery. There are no published efficacy data for either of them; they appear as strategies to be tried out in training, not as validated protocols.

Women. An audit of sex representation in research on ergogenic supplements found that only 4% of bicarbonate studies had been carried out with female participants (Smith et al., 2022). A small later meta-analysis suggests that the standard doses of 0.2 to 0.3 g/kg raise buffering capacity and improve performance in women as well (Saunders et al., 2022). The honest conclusion is not that it works the same: it is that the basis for saying so is among the narrowest in the catalog.

The protocols with backing use between 0.2 and 0.4 grams per kilo of body weight — 0.3 g/kg, the midpoint of the range, is about 21 grams for someone of 70 kg — taken 120 to 150 minutes before the effort and together with a carbohydrate-rich meal.

Three details of those protocols change the experience a great deal: splitting the intake over 30 to 60 minutes instead of swallowing it all at once, taking it with food, and drinking around 10 milliliters of fluid per kilo of body weight, because without enough fluid the bicarbonate drags water into the intestine.

There is also a multi-day protocol, with the daily dose split between meals for up to five days before competing. In the studies, that protocol is associated with fewer digestive problems on competition day.

The AIS is explicit on one point: the protocol is tried out in training, several times, before being taken to a competition.

What is not clear is what follows from that variability. It has been proposed that the decisive thing is not the dose swallowed but how much bicarbonate rises relative to each person’s own value. A 2022 meta-analysis found more effect when that rise exceeded 4 mmol/L than when it fell short of it; the one by Grgic et al. (2020) found no significant linear relationship between the rise and performance. They are, moreover, averages across studies, not a rule anyone can apply to themselves.

The authors conclude that it has to be individualized. Doing that for real is another matter: it requires measuring blood bicarbonate several times over several sessions, with a laboratory or a blood gas analyzer. And as to whether that effort translates into performance, the evidence is still not conclusive.

The dose-response study that underpins the individualization problem is Jones et al. (2016): 16 healthy men, three randomized single-dose sessions of 0.1, 0.2 and 0.3 g/kg. The increases in blood bicarbonate were +2.0 to 5, +5.1 to 8.1 and +6.0 to 12.3 mmol/L, respectively; the times to peak, 30 to 150, 40 to 165 and 75 to 180 minutes. It was an open-label study with no placebo, a design suited to measuring blood analytes, but not to measuring performance; its ranges describe kinetics, not efficacy.

What happens after that is still an open question and it is best said that way. The AIS holds that there is some evidence that matching the competition to the individual peak of buffering capacity can improve performance, and it cites three works to that effect (Heibel et al., 2018; Boegman et al., 2020; Oliveira et al., 2020). It also acknowledges the cost: it requires measuring blood bicarbonate repeatedly over several sessions, with a blood gas analyzer or access to a laboratory, and it warns that the method’s performance falls off if the athlete has a long warm-up. For anyone who cannot individualize, the AIS itself describes an alternative: the high end of the dose range, between 300 and 400 mg/kg, two to three hours before, which raises buffering capacity to presumably ergogenic levels with an effect that holds for three or four hours.

As for the idea that there is a personal rise threshold above which the supplement works, it is worth being precise about where it comes from. The Bayesian review by Farias de Oliveira et al. (2022) classified mean bicarbonate rises into three bands — ≤4, 4-6 and >6 mmol/L — and found a greater effect on exercise in the medium and large bands than in the small one; for the medium band, β = 0.16 (95% CrI 0.02–0.32).

Three qualifications that tend to get lost when it is cited. The first: those labels classify the rise in bicarbonate, not the size of the improvement; the overall effect of the supplement in that same work is ES₀.₅ = 0.17 (95% CrI 0.12–0.21), small by any convention. The second: it is a meta-regression with a group-level moderator, which compares means across studies and not individuals within a study, so it cannot support “if you get to 6, you improve a lot”. The third: it is not a meta-analysis exclusive to bicarbonate, but of extracellular buffers as a whole, with 158 of its 189 articles devoted to bicarbonate.

The AIS structures the introduction of the supplement in four phases with the support of a sports dietitian: preparation and an audit of previous supplementation; buffering trials at rest, quantifying digestive symptoms every 30 minutes; trials during high-intensity training sessions; and lastly, use in competition, starting with a low-priority event.

Nausea, abdominal pain, bloating, diarrhea and vomiting. It is the factor that decides whether bicarbonate serves you or ruins your day, and it is not a rarity.

In the continuous running trials, digestive symptoms appeared in 29.5% of those who took bicarbonate versus 2.6% with placebo, and 8.7% dropped out of the study for that reason, versus 1.6% in the placebo group.

It is not just discomfort. When it appears in competition, the studies describe that it can erase that 2 or 3% improvement and leave whoever takes it below where they would have been had they taken nothing.

The strategies for reducing it have backing — capsules instead of dissolved powder, enteric-coated capsules, hydrogel systems, lower doses or the multi-day protocol — but none of them eliminates it for everyone.

A large load of alkaline salt in the stomach does two things at once: it reacts with gastric acid releasing carbon dioxide, and it drags water into the intestine by difference in concentration. Hence the bloating, the belching and the diarrhea, and hence too that the peak of discomfort arrives around ninety minutes in, just before competing if it was taken at the standard time.

The strategies that reduce them are measured, not assumed: capsules avoid contact between the solution and the mucosa; splitting the intake reduces the concentration peak; the enteric coating moves dissolution to the intestine; drinking enough fluid prevents osmotic diarrhea; and the multi-day protocol spreads out the load. Hydrogel systems show a low incidence of symptoms, but the evidence is scarce and has been generated in a single sport.

The work that set the mitigation protocol is Carr et al. (2011), carried out at the AIS itself: it systematically compared the time taken to consume the load, split over 30 to 60 minutes; the form of administration, flavored powder versus capsules; and different amounts of accompanying food and fluid. The combination that best reconciled a rise in blood bicarbonate with a lower incidence of symptoms was administration in capsules, in a split intake, started 120 to 150 minutes before and accompanied by carbohydrate-rich food and fluid.

Two practical consequences follow from the kinetics of the symptoms, which reach their maximum around 90 minutes after intake. The first: bringing the intake forward to the two- or three-hour end gives room for the discomfort to resolve before the start. The second: in sports with weight categories, the volume of fluid needed to prevent osmotic diarrhea introduces a weight gain that also has to be counted.

The effect can even be reversed: Grgic et al. (2020) point out expressly that in someone who suffers gastric discomfort bicarbonate can turn out to be ergolytic, that is, to worsen performance instead of improving it.

To quantify the symptoms there is a validated visual analog scale questionnaire with published test-retest reliability (Gaskell et al., 2019). The AIS notes that applying it systematically frequently uncovers correctable problems in the individual protocol.

On the topical route, the three available trials do not agree: benefit both oral and topical in a battery of team-sport tests (Gurton et al., 2023); benefit oral and not topical in simulated soccer (Gurton et al., 2024); and no changes in buffering capacity or in performance (McKay et al., 2020).

Outside sport, the uses of bicarbonate backed by clinical trial are medical and supervised: correcting the acidity of the blood in some kidney diseases and alkalinizing the urine in specific situations. They are treatments that a professional prescribes and monitors.

As a home antacid it is very widespread and there are barely any trials that back it. What is documented are its risks, and the main one goes unnoticed easily: a sporting dose of 21 grams supplies about 5.75 grams of sodium, almost three times the daily maximum the WHO recommends.

Advanced chronic kidney disease can produce metabolic acidosis, and correcting it with oral bicarbonate is established clinical practice. It is a treatment with an indication, a dose and laboratory follow-up, not a supplementation.

The evidence on whether it also slows the deterioration of the kidney is discordant. A 2009 trial in patients with advanced kidney disease found a much slower progression and fewer entries into dialysis; it was open-label and without a placebo, something its own authors declare. A later trial, multicenter and double-blind, measured physical function and not renal progression as its primary outcome: it found no benefit in either of the two, and it did find more adverse effects. When two studies disagree and only one is blinded, the blinded one is the one that carries the weight.

In all cases this is a prescribed and monitored indication; the prudent thing is not to modify it on your own.

Outside that indication, domestic use as an antacid is very widespread and very little studied. Anyone who takes it daily is adding salt to their diet without counting it, and that load weighs especially in hypertension, heart failure and kidney disease, three of the most frequent reasons for taking antacids.

The disagreement about renal progression pits two trials against each other that are not comparable in methodological quality.

de Brito-Ashurst et al. (2009), in J Am Soc Nephrol (PMID 19608703), was an open-label trial with no placebo, made explicit by its authors. It described a much slower progression of renal deterioration and fewer entries into dialysis in the bicarbonate group. The origin of the conflict of interest declaration that circulates associated with this work could not be identified in the original journal, and it is therefore not reproduced here.

BiCARB (2020), in BMC Medicine (PMID 32268897; doi:10.1186/s12916-020-01542-9), was multicenter, randomized, double-blind and placebo-controlled, with 300 elderly participants. It found no difference in physical function, which was its primary outcome, nor in renal function measured as a secondary outcome; it recorded more adverse events in the bicarbonate arm (457 versus 400), worse quality of life measured with the EQ-5D-3L and a higher cost; placebo dominated bicarbonate in all the sensitivity analyses.

An open-label positive result and a blinded negative result are not averaged. This does not invalidate the use of bicarbonate to correct acidosis in kidney disease, which is a different and established indication: what is left unsupported is the added expectation of slowing its progression. In all cases this is a prescribed indication, with a dose and laboratory follow-up; the prudent thing is not to modify it on your own.

The arithmetic of sodium. An ergogenic dose of 0.3 g/kg for 70 kg is 21 grams of sodium bicarbonate. With a molar mass of 84 g/mol, that is equivalent to 250 millimoles, and therefore to 250 millimoles of sodium: 5.75 grams. The daily limit the WHO recommends is 2 grams, so that a single intake supplies 2.875 times that figure.

Milk-alkali syndrome. When the alkali source is combined with calcium, the picture described includes hypercalcemia, alkalosis and renal deterioration, and it has been pointed to as one of the most frequent causes of hospital hypercalcemia (Medarov, 2009; doi:10.4065/84.3.261). Sodium bicarbonate on its own produces above all alkalosis and hypernatremia, not hypercalcemia; the risk appears with the combination.

They are two of the buffers with backing and they work in different places: beta-alanine builds carnosine inside the fiber over weeks; bicarbonate acts outside it and within hours. It is also one of the few supplement combinations that the reviews point to as probably useful.

The data are scarce. The most cited work, from 2013, compared the two separately and together in high-intensity cycling: around 7% and 8% each, 14% combined. A single study, and a small one; the later review keeps the conclusion in the realm of the promising.

Sodium bicarbonate does not appear on the World Anti-Doping Agency’s list of prohibited substances. Each federation and each national body also has its own rules.

It does have one little-known practical consequence: taking it alters the pH of the urine, which can take several hours to return to the values that anti-doping controls accept. It is not a positive, but it can upset the routine of an athlete who happens to be tested that day.

The relevant risk is not the bicarbonate, but the packaging: any supplement can be contaminated during its manufacture, and that is why anti-doping bodies recommend products with batch testing. The risk depends on the number of ingredients and on the traceability of the batch: with a single-ingredient product it is low, and it grows as the formula takes on components.

WADA’s List of Prohibited Substances and Methods for 2026 was approved on September 11, 2025 and came into force on January 1, 2026. Sodium bicarbonate does not appear on it.

The alteration of urinary pH after intake does not constitute a violation, but it can delay by several hours the obtaining of a sample within the values that the testing authorities accept. Under the strict liability rule of the World Anti-Doping Code, the athlete is answerable for everything that enters their body, which shifts the real risk onto the specific product and the traceability of its batch.

How much of the effect can be put down to expectation. In 2016 a fake bicarbonate was prepared with carbonated water and salt, designed to imitate the stomach heaviness of the real product; on top of that, participants were told it would improve their performance. They lasted 9.5% longer to exhaustion than with the neutral placebo. There were eight people and the effect size was small, but the figure exceeds the average attributed to real bicarbonate. The society’s position stand admits that part of the effect observed may be placebo.

What happens in the brain. Nothing verified: there are no randomized placebo-controlled trials that have measured memory, attention or executive function. What is sometimes presented as a cognitive effect is in reality a reduction in perceived exertion, which is neither the same thing nor does it appear consistently.

Who responds and who does not. There is no practical way of knowing in advance, other than measuring the blood repeatedly, something that requires equipment and a laboratory.

The expectation experiment, with its numbers. Eight recreationally active men cycled at 100% of their peak power to exhaustion, 60 minutes after taking a neutral placebo or a fake bicarbonate made up of sodium chloride and carbonated water, formulated to reproduce the fullness and the gastric discomfort of the real product, accompanied by a script that announced an improvement in performance. Time to exhaustion: 461 ± 148 seconds versus 421 ± 150, 9.5% more (p=0.048; d=0.3). A tiny sample and a small effect size; even so, the point magnitude exceeds the average attributed to the active compound (Higgins & Shabir, 2016; PMID 26863442).

The blinding. Gurton et al. (2022) asked twelve athletes what they thought they had taken, at three moments and with two forms of administration. With dissolved bicarbonate the correct guesses were 42% after consumption and 67% before exercise; with capsules, 0% and 25%. A second reading of Table 2 adds a third measurement, after exercise: 75% in solution and 33% in capsules, figures that appear in only one of the two verifications. The authors do not conclude that the blinding broke: there were no significant differences in allocation between bicarbonate and its equivalent placebo at any moment. What they do point out is that identification of dissolved bicarbonate exceeded the 50% to be expected by chance, and that part of the ergogenic benefit could therefore be due to participant bias. With twelve participants each person is worth 8.3 percentage points, and the figure should not travel without that fact. In an earlier tasting experiment, the masking agent changed the result a great deal: with cold orange 44% got it right; with cold blackcurrant, 69%. (Eur J Appl Physiol 2022;122(12):2555-2563; doi:10.1007/s00421-022-05031-0; PMID 36053364. No external funding and no declared conflicts of interest.)

Cognition: a complete gap. No randomized, double-blind, placebo-controlled trial has been located that has measured memory, attention, processing speed or executive function after sodium bicarbonate intake. What is on occasion presented as a cognitive effect is the reduction in perceived exertion, a different outcome, measured with subjective scales and appearing inconsistently: Artioli’s own trial found no differences in that variable. The gap is declared; it is not filled in with plausible mechanisms.

Who writes the recommendations

A good part of the sporting framework for bicarbonate rests on the position stand of the International Society of Sports Nutrition (Grgic et al., 2021). In the conflict of interest declaration published with that document, one of the authors was listed as chief executive of the society itself and two as members of the scientific advisory board of one and the same sports supplement manufacturer, specifying that when the review was prepared and published that company did not produce bicarbonate supplements. The document is also published in the society’s own journal.

The precedent is documented in other position stands by the same body: the one on creatine, published in 2017, recorded in its conflict of interest declaration that some of its authors then held consultancies or worked in organizations that received funding from manufacturers of the product they were pronouncing on.

None of this invalidates the content, which is the part that has to be said clearly: they are peer-reviewed documents with verifiable data. It does mean they have to be read as what they are, an expert consensus with declared ties, and not as a neutral source.

The sodium load is the first reason not to take it on your own. The prudent thing is to discuss it with a professional in cases of hypertension, heart failure, kidney disease or any indication to restrict salt, and also during pregnancy and breastfeeding, where information is lacking.

All the protocols described come from studies carried out in adults. No evidence has been located on the use of bicarbonate in minors, nor any specific protocols for them.

There is a second reason, less obvious. By modifying the acidity of the stomach and of the urine, bicarbonate alters the absorption or the elimination of several medications; some official prescribing documents name it expressly. Anyone who takes medication regularly has a conversation there that is worth having with their doctor or their pharmacist.

A note about doses: the serious published cases come from intakes far above the sporting ones, sustained for weeks or massive all at once. Above the range the studies work with there are no data of additional benefit, and digestive symptoms increase with the dose.

Interactions with medications. The mechanism is twofold: by neutralizing gastric acid, bicarbonate hinders the dissolution of some drugs that need an acidic medium in order to be absorbed; by alkalinizing the urine, it also modifies the rate at which the kidney eliminates others. Several official prescribing documents record it, and some name sodium bicarbonate expressly: among them those of an oral antibiotic, lithium and prescription amphetamines. The separation intervals vary according to the drug and appear in each prescribing document; the reliable answer is with the doctor or the pharmacist, not in a general table.

Overdose. A structured review of 78 published cases describes the picture: severe metabolic alkalosis, elevated blood sodium and disturbances of other electrolytes, with manifestations ranging from weakness and nausea to seizures and arrhythmias. The cases follow two patterns — sustained chronic intake and massive acute intake — and in both the amounts involved are several times the dose used in sport.

Gastric rupture. It is the most serious complication described, but also the rarest: one review identified 36 cases published in the world literature, of which 10 were associated with sodium bicarbonate. Among the published cases there is none associated with the doses used in sport: the cases described are concentrated in massive intakes, close to half of them in the context of an eating disorder. Eating disorders have treatment, and seeking professional help is the prudent thing both for someone who suffers from one and for someone who recognizes one in a person close to them.

Interactions: what is documented and in which source. At least three official FDA prescribing documents name sodium bicarbonate expressly. The one for cefpodoxime proxetil describes that the concomitant administration of high doses of antacids, among them sodium bicarbonate, reduces peak plasma concentrations and the extent of absorption. The one for lithium carbonate includes it among the alkalinizing agents that lower serum lithium concentration by increasing its urinary excretion; lithium is a drug with a narrow therapeutic margin, so that drop amounts to a loss of treatment efficacy. The one for prescription amphetamines describes the twofold route: gastrointestinal alkalinizers increase their absorption and urinary ones reduce their elimination, both raising blood concentrations.

The prescribing documents of a second group of drugs describe that their absorption decreases with agents that raise gastric pH, without naming bicarbonate: atazanavir, dasatinib, ketoconazole, itraconazole and levothyroxine. In the case of mycophenolate mofetil the document records the interaction with magnesium and aluminum antacids and with proton pump inhibitors, not with bicarbonate, which is why it is presented here as an effect of the agents that raise gastric pH and is not attributed to it by name.

Each prescribing document specifies its own separation interval between intakes, and they are not reproduced here: they are prescribing data whose reading is the responsibility of the doctor or the pharmacist. None of the above is a reason to interrupt or modify a treatment on your own.

Declared gaps. No FDA or AEMPS prescribing document for oral sodium bicarbonate with an interactions section has been located. Nor has an admissible primary source been located for the magnitude by which bicarbonate reduces the absorption of oral iron: the interaction is plausible by mechanism, but the figure in circulation comes from a commercial aggregator and is not reproduced.

Toxicity. A structured review of 78 published cases distinguishes two patterns: sustained chronic intake, which presents with metabolic alkalosis and electrolyte disturbances, and massive acute intake, frequently associated with gastric rupture and with high mortality (Inokuchi et al., 2026; PMID 41650758). Its index case describes a healthy man who consumed 60 grams a day for two months, with a much higher final intake, and presented with severe metabolic alkalosis and a serum sodium of 162 mEq/L. Earlier series also document hypokalemia, hypoxia and, as possible presentations, seizures, arrhythmias and cardiorespiratory arrest (Thomas et al., 1999; PMID 9950389). The authors of that review also point to a shift in the reasons for consumption: from dyspepsia toward uses without backing, and they attribute it expressly to unsupported health claims in circulation.

Gastric rupture, with the figures. The reference review identified 36 published cases: 69% in women, mortality of 41.6%, and sodium bicarbonate implicated in 10 of them. 47.2% correlated with an eating disorder (Han et al., 2022; doi:10.1186/s40337-022-00677-9; PMID 36348449). The most frequent locations were the lesser curvature and the anterior wall of the stomach. Among the published cases there is none associated with the doses used in sport.