Sodium bicarbonate

What it is

Sodium bicarbonate is a simple sodium salt of carbonic acid, familiar as baking soda. Swallowed, it reacts with stomach acid to make carbon dioxide, water and salt, which is why it is sold as an antacid; a US Drug Facts label for a 650 mg tablet gives its purpose as antacid. Absorbed, it adds alkali and sodium to the blood, which is the basis of its three other uses: correcting the acidosis of advanced kidney disease, making urine alkaline, and loading the blood with buffer before high-intensity exercise. The sodium matters: the same label states each 650 mg tablet contains sodium 178 mg.

What the research says

The picture divides sharply by use. For acidosis in chronic kidney disease the literature is genuinely split: a 134-patient trial found far less rapid decline and far less progression to end-stage disease, while the larger, blinded BiCARB trial in 300 older people found no benefit on physical or kidney function, more adverse events and worse measured quality of life. In cardiac arrest the evidence is negative: a 2026 randomised trial in 779 analysed patients found no difference in return of circulation, and a meta-analysis of observational data found none either. In critical-illness acidaemia the BICAR-ICU trial missed its primary outcome overall. For high-intensity exercise there is a real but small effect, positive in cycling time trials and negligible in running once dropouts and publication bias are allowed for, with frequent gut upset. Harms across uses are metabolic alkalosis, high blood sodium, diarrhoea and bloating, and rare gastric rupture after large doses on a full stomach.

Evidence grade: Disputed.

How it works

Drug class: Alkalinising agent and systemic/gastric antacid (sodium salt of carbonic acid)

Sodium bicarbonate is an alkali. Absorbed into the blood it raises the serum bicarbonate concentration, which is how it corrects the metabolic acidosis of chronic kidney disease: a 2025 meta-analysis of 20 randomised and 2 non-randomised trials in 2,932 patients found serum bicarbonate rose by a mean difference of 2.59, although with extreme heterogeneity between the trials. Its acid-neutralising action in the stomach and its effect on urine pH are reported under the separate uses below, each with its own source. (Source 1)

What it is used for

• The approved antacid use rests on the chemistry and on placebo-controlled trials of antacids as a class rather than on trials of sodium bicarbonate tablets specifically. In a 565-patient placebo-controlled trial of self-treated heartburn, the median proportion of a patient's episodes relieved was 62% with an unnamed antacid of 11 mEq acid-neutralising capacity against 41% with placebo - so it speaks to antacids as a class, not to sodium bicarbonate. The sodium load and a rare risk of gastric rupture are the reasons it is the least favoured antacid. Evidence: limited. (Source 2)
• It reliably raises serum bicarbonate. Whether that changes anything a patient would notice is disputed: one 2-year open trial against standard care in 134 adults reported a large reduction in progression to end-stage kidney disease, while the blinded placebo-controlled BiCARB trial in 300 older people - with primary outcome data for 187 - found no benefit on physical function or filtration rate and more adverse events. That is a failure to show benefit rather than a demonstration of equivalence. Meta-analysis found no significant improvement in estimated glomerular filtration rate. A Cochrane review of sodium bicarbonate for acute kidney injury, a different setting, found no randomised trial met its criteria at all. Evidence: disputed. (Source 3)
• It does what it is meant to do chemically. In an uncontrolled before-and-after pilot in 33 women with acidic urine, four weeks of oral sodium bicarbonate raised urine pH from 5.31 to 7.2 while blood pH did not change significantly. A systematic review of oral regimens for high-dose methotrexate found no significant differences against intravenous protocols on most outcomes, with nearly three times the odds of diarrhoea on the oral route. The evidence is small studies and mostly non-randomised comparisons. Evidence: limited. (Source 4)
• A 2026 double-blind randomised trial in 21 Danish hospitals found no difference in return of spontaneous circulation (39% against 37%, risk ratio 1.05, 95% CI 0.88 to 1.24, P = .62) and more alkalosis and high blood sodium. A separate meta-analysis of six observational studies in 18,406 patients found no improvement in return of circulation (OR 1.185, 95% CI 0.680 to 2.065) or survival to discharge (OR 0.296, 95% CI 0.066 to 1.323). In the BICAR-ICU trial of severe acidaemia the primary composite outcome was not improved (71% against 66%, absolute difference -5.5%, 95% CI -15.2 to 4.2, p=0.24), although a prespecified acute kidney injury stratum showed better day-28 survival (37% against 54%, p=0.0283). Each figure here comes from the trial or review named beside it. Evidence: not-supported. (Source 5)
• A meta-analysis of 18 cycling time-trial studies found small improvements in mean power output and time. A 2025 meta-analysis of running trials found the effect negligible and not statistically significant once gut-related dropouts and publication bias were accounted for, and gut symptoms were far more common than on placebo. Evidence: disputed. (Source 6)

Interactions

• Sodium in the diet, and a sodium-restricted diet (label): Every tablet brings sodium with it, which is why the label tells people on a sodium-restricted diet to ask a doctor first. The labelled tablet contains sodium 178 mg. (Source 7)
• Dietary sodium limits at ergogenic doses (clinical trial): At the doses used to improve exercise performance, the sodium delivered exceeds the whole recommended daily sodium limit on its own: 300 mg/kg in a 70 kg adult is about 21 g of sodium bicarbonate, which at 27% sodium is roughly 5.7 g of sodium against a Dietary Reference Intakes upper limit of 2,300 mg a day. Limit: The trial this comes from was not uniformly null on the circulation, which its title's word "negligible" might suggest: systolic blood pressure was unaffected (P=.38), but diastolic pressure rose by 3.3 +/- 1.1 mmHg (P=.01) and heart rate by 10.1 +/- 2.4 beats a minute (P=.002). "Negligible" is the authors' own characterisation. Eleven participants. (Source 8)
• A large meal or binge eating (case reports): Taken after a large meal, the carbon dioxide it releases can over-distend a full stomach; published cases describe spontaneous gastric rupture in this setting. (Source 9)
• High-dose methotrexate (clinical trial): It is used deliberately with methotrexate, to keep urine alkaline so the drug does not damage the kidneys; this is a wanted interaction rather than a hazard. (Source 10)
• Prescription medicines generally, through the rise in stomach pH (label): The label warns that antacids may interact with prescription medicines; the pH-dependent interactions documented for other antacids, such as riociguat and oral antibiotics, arise by the same mechanism. (Source 7)
• Alcohol (case reports): UNTESTED, not tested and found safe. We found no study of a sodium bicarbonate and alcohol interaction in this search. The documented food-related hazard is a large meal rather than alcohol: the gastric rupture series links cases to binge eating and to gas release from sodium bicarbonate, and says nothing about drinking either way. Limit: A gap in the literature, not a negative result. Nobody has run the study, so there is no evidence either that alcohol matters here or that it does not. The case series is a series of 36 published cases and cannot give a rate per user for anything. (Source 11)

Stopping it

• There is no withdrawal syndrome described for sodium bicarbonate, and no deprescribing trial. For the antacid use, the label itself sets a two-week limit and a stopping rule. (Source 7)
• For the kidney-disease use, the strongest evidence on continuing it is the BiCARB trial's own conclusion, which weighed the lack of benefit against the increase in adverse events and the cost. (Source 12)
• For cardiac arrest, the 2026 randomised trial's authors stated the implication for practice directly, that routine use is not supported. (Source 13)

What goes wrong

In that trial alkalosis and high blood sodium after cardiac arrest were more common in the sodium bicarbonate group. (Source 5)

• Randomized trial, High certainty.
• Size: 779 patients analysed.
• Who: Adults with in-hospital cardiac arrest.
• How long: After the arrest.
• Result: The abstract reports the direction without rates: alkalosis and hypernatraemia were more common with sodium bicarbonate.
• Funding: not stated in the abstract.

Alkalosis and hypernatremia after cardiac arrest were more common in the sodium bicarbonate group.

In the same running meta-analysis, gut symptoms and gut-related withdrawals were significantly more frequent on sodium bicarbonate than on placebo, at odds ratios of 5.9 and 2.9. (Source 14)

• Meta-analysis, Low certainty.
• Size: 11 studies, 126 participants.
• Who: Runners in blinded placebo-controlled crossover trials.
• How long: Single doses.
• Result: Gastrointestinal symptoms 29.5% on sodium bicarbonate against 2.6% on placebo (odds ratio 5.9, 95% CI 1.8 to 18.8, p = 0.003, low certainty); gut-related study withdrawal 8.7% against 1.6% (odds ratio 2.9, 95% CI 1.0 to 8.1, p = 0.049, moderate certainty). These are the review's own published effect measures.
• Funding: not stated.

Limit of this finding: This entry previously said gut symptoms were "about eleven times" and withdrawals "five times" more frequent. Those were ratios of the review's raw percentages, not figures the review publishes, and they contradicted the review's own odds ratios of 5.9 and 2.9, so we have dropped them and quote the odds ratios instead. The review's own numbers are also internally inconsistent: a two-by-two table on 29.5% against 2.6% gives an odds ratio of about 15.7, not 5.9, and one on 8.7% against 1.6% gives about 5.9, not 2.9 - the two odds ratios look as though they may have been transposed between the two lines. We have reproduced the review as published and recomputed nothing. What is not in doubt is the direction and the significance: gut upset was much more common on sodium bicarbonate.

GI symptoms occurred more frequently with SB than placebo (29.5% vs. 2.6%; odds ratio = 5.9; p = 0.003; low certainty), as did GI-related study withdrawal (8.7% vs. 1.6%; odds ratio = 2.9; p = 0.049; moderate certainty).

In a double-blind crossover trial, an ergogenic dose of sodium bicarbonate caused diarrhoea in 10 of 11 trained participants and substantial gut distress. (Source 15)

• Randomized trial, Very low certainty.
• Size: 11 endurance-trained men and women.
• Who: Endurance-trained adults exercising 4.6 +/- 0.4 sessions a week.
• How long: Single 300 mg/kg dose versus placebo on two occasions.
• Result: Global gut distress rating 5.1 +/- 0.5 versus 0.5 +/- 0.2 on a 0 to 10 scale (P < .0001); diarrhoea in 10 of 11 (91%), bloating and thirst in 64%, nausea in 45%.
• Funding: Research Support, Non-U.S. Gov't per the record.

Limit of this finding: The percentages come from very small counts and a percentage on a health screen reads larger than it is: with 11 participants, 91% is 10 people, 64% is 7 and 45% is 5. The source also prints a tense slip inside the quoted sentence, "64% experience bloating and thirst", which we have reproduced rather than tidied.

Furthermore, 10 of the 11 subjects (91%) experienced diarrhea, 64% experience bloating and thirst, and 45% experienced nausea after bicarbonate loading.

A review of published cases found 36 spontaneous gastric ruptures, 10 of them associated with sodium bicarbonate ingestion, with a mortality of 41.6% across the series. (Source 11)

• Case series, Very low certainty.
• Size: 36 identified cases, plus 2 new cases reported by the authors.
• Who: Adults with spontaneous gastric rupture; about half the cases linked to eating disorders, 69% female.
• How long: Cases accrued across the published literature.
• Result: Sodium bicarbonate ingestion associated with 10 of the 36 cases; mortality 41.6%; commonest rupture sites the lesser curvature (36.1%) and anterior wall (33.3%). Case series cannot give an incidence or a rate per user.
• Funding: not stated.

Amongst the 36 cases, sodium bicarbonate ingestion was associated with 10 cases.

The US label for this sodium bicarbonate antacid carries a California Proposition 65 warning that the product can expose the user to lead and mercury. (Source 16)

• Official position, Certainty not rated.
• Size: Not applicable - a labelling disclosure.
• Who: Anyone using the over-the-counter 650 mg antacid tablets, which the same label allows up to 24 a day for adults under 60.
• How long: Position dated 25 September 2026 (SPL version 7 published date)
• Result: No quantity is given. The label prints the warning with no lead or mercury content figure and no measured exposure anywhere, so it discloses that an exposure exists without saying how large it is. The warning sits in the SPL's unclassified package-insert section, outside the Drug Facts panel.
• Funding: Not applicable.

Limit of this finding: A Proposition 65 warning is a California labelling requirement, not a dose or a risk estimate: it tells you an exposure above California's threshold exists, and nothing about how much lead or mercury is in a tablet, because the label does not say. It sits in the package-insert text outside the Drug Facts panel, which is why the earlier excerpt of this label did not carry it. We found no study measuring heavy metals in sodium bicarbonate antacid tablets.

WARNING: THIS PRODUCT CAN EXPOSE YOU TO CHEMICALS INCLUDING LEAD AND MERCURY, WHICH ARE KNOWN TO THE STATE OF CALIFORNIA TO CAUSE BIRTH DEFECTS OR OTHER REPRODUCTIVE HARM. FOR MORE INFORMATION GO TO WWW.P65WARNINGS.CA.GOV.

What the evidence supports

In a 2-year open trial in 134 adults with advanced chronic kidney disease and low serum bicarbonate, far fewer people in the bicarbonate group progressed rapidly or reached end-stage kidney disease than in the standard-care group, and their kidney function declined more slowly. (Source 3)

• Randomized trial, Low certainty.
• Size: 134 adults.
• Who: Adults with creatinine clearance 15 to 30 ml/min per 1.73 m2 and serum bicarbonate 16 to 20 mmol/L.
• How long: 2 years.
• Result: Every pair below is given as BICARBONATE versus STANDARD CARE. Rapid progression 9% on bicarbonate versus 45% on standard care (relative risk 0.15, 95% CI 0.06 to 0.40, P < 0.0001). End-stage kidney disease 6.5% on bicarbonate versus 33% on standard care (relative risk 0.13, 95% CI 0.04 to 0.40, P < 0.001), an absolute difference of about 26 percentage points. Decline in creatinine clearance was SLOWER on bicarbonate: the source prints that pair the other way round, as 5.93 ml/min 1.73 m2 in the CONTROL group versus 1.88 in the BICARBONATE group. The comparator was standard care, not placebo, and the trial was not blinded.
• Funding: Research Support, Non-U.S. Gov't per the record.

Limit of this finding: Two things a reader should know. First, the creatinine-clearance pair is printed control-first in the source while the other two pairs are intervention-first; read as the source's own sentence makes clear, bicarbonate was the slower decline, so bicarbonate is 1.88 and standard care 5.93. Second, the relative risks the paper prints do not equal the ratios of the percentages it prints beside them: 9% against 45% is a risk ratio of 0.20, not the printed 0.15, and 6.5% against 33% is 0.20, not the printed 0.13. Both printed relative risks are internally consistent with their own confidence intervals, so the discrepancy is the source's own and we have not altered its figures - but the percentages and the relative risks cannot both be right. The direction is not in doubt: bicarbonate did better on every measure. The size of the benefit is less certain than the printed numbers suggest. The trial was also open against standard care rather than placebo, which inflates benefit on outcomes that depend on clinician decisions.

Patients supplemented with bicarbonate were significantly less likely to experience rapid progression (9 versus 45%; relative risk 0.15; 95% confidence interval 0.06 to 0.40; P < 0.0001).

The same meta-analysis found no significant excess of death or prolonged hospitalisation, gastrointestinal disorders or worsening oedema with sodium bicarbonate, though the confidence intervals were wide. (Source 1)

• Meta-analysis, Very low certainty.
• Size: 22 studies, 2,932 patients.
• Who: Patients with chronic kidney disease and metabolic acidosis.
• How long: Searched to May 2024.
• Result: Death or prolonged hospitalisation RR 1.05 (95% CI 0.84 to 1.32, P = 0.66); gastrointestinal disorders RR 1.64 (95% CI 0.35 to 7.66, P = 0.53); worsening oedema RR 1.26 (95% CI 0.94 to 1.68, P = 0.12)
• Funding: not stated.

no significant association was found in the outcomes of death/prolonged hospitalization (RR: 1.05, 95%CI: 0.84-1.32; P = 0.66; I 2 = 0%), or gastrointestinal disorders (RR: 1.64, 95%CI: 0.35-7.66; P = 0.53; I 2 = 76%), or worsening edema (RR: 1.26, 95%CI: 0.94-1.68; P = 0.12; I 2 = 37%) when compared to control

A meta-analysis of 18 cycling time-trial studies found small improvements in mean power output and in performance time with sodium bicarbonate against placebo. (Source 6)

• Meta-analysis, Low certainty.
• Size: 18 studies; 182 observations for mean power output and 201 for time performance.
• Who: Cyclists in randomised placebo-controlled studies.
• How long: Acute single doses.
• Result: Mean power output standardised mean difference 0.42 (95% CI 0.21 to 0.63, P = 0.001); performance time SMD 0.22 (95% CI 0.02 to 0.43, P = 0.03); heterogeneity I squared 0% for both.
• Funding: not stated.

Overall, when compared to placebo, the NaHCO3 ingestion improved the Wmean (SMD: 0.42; 95% CI: 0.21-0.63; P = 0.001) and performance time (SMD: 0.22; 95% CI: 0.02-0.43; P = 0.03).

A meta-analysis of 14 randomised trials in 2,110 people with chronic kidney disease found sodium bicarbonate did not raise systolic blood pressure. (Source 17)

• Meta-analysis, Moderate certainty.
• Size: 14 randomised controlled trials, 2,110 individuals.
• Who: Adults with chronic kidney disease stages G1 to G5 not on dialysis; mean age 60, mean systolic blood pressure 136 mm Hg, mean eGFR 38 ml/min.
• How long: Median follow-up 27 weeks.
• Result: No influence on systolic blood pressure, consistent across dose and duration strata; no significant increase in antihypertensive or diuretic use, and a greater decrease in antihypertensive use with bicarbonate.
• Funding: Research Support, Non-U.S. Gov't per the record.

Limit of this finding: Two notes. The source prints its follow-up as "median follow-up 27 [interquartile range 97] weeks" - an interquartile range of 97 weeks around a median of 27, which cannot be right as a dispersion statistic. It is printed that way in the published Results section and we have not altered it. And the abstract states the null for systolic blood pressure in words only: it gives no point estimate and no confidence interval, so there is no effect size to report here and none should be supplied later. The authors' own certainty rating for the finding is moderate.

Meta-analysis suggested that sodium bicarbonate did not influence systolic BP in individuals with CKD stage G1-5.

In an uncontrolled before-and-after pilot in 33 women with lower urinary tract symptoms and acidic urine, four weeks of oral sodium bicarbonate raised urine pH, while blood pH did not change significantly. (Source 18)

• Case series, Very low certainty.
• Size: 33 female patients.
• Who: Women with lower urinary tract symptoms and a urine pH below 6; single-arm, no control and no placebo group.
• How long: 4 weeks, 2 x 4 g a day - an 8 g daily sodium bicarbonate load.
• Result: Urine pH rose from 5.31 +/- 0.52 to 7.2 +/- 0.66 (p < 0.001). Blood pH did not change significantly (p = 0.14). Daily frequency, nocturia, urgency, urge incontinence and all symptom scores fell. There was no control arm, so none of the symptom changes can be separated from the natural course or from being in a study.
• Funding: not stated.

Limit of this finding: Two cautions. First, the published blood pH figures are not physiologically possible as printed: "7.369 +/- 0.33 to 7.384 +/- 0.28" implies blood pH ranging roughly from 7.04 to 7.70, which nobody survives at the extremes. They are almost certainly misprints for +/- 0.033 and +/- 0.028. The figures are identical in the Europe PMC record and in the NCBI record, so the error is the publisher's, not a transcription slip, and we have quoted it as printed rather than correct someone else's paper. Do not read those standard deviations as real. The part of this result that stands is the direction: urine pH rose and blood pH did not change significantly (p = 0.14). Second, this was an uncontrolled single-arm pilot in 33 women with no placebo group, which the design label now reflects: it measured symptoms before and after, not blood levels, so the earlier "pharmacokinetic" label was wrong.

A significant increase was detected in urine pH values measured after treatment (5.31 ± 0.52 to 7.2 ± 0.66, p < 0.001), but not in blood pH values (7.369 ± 0.33 to 7.384 ± 0.28, p = 0.14).

What the evidence does not support

In the BiCARB trial, a double-blind placebo-controlled trial that randomised 300 older people with advanced kidney disease but had primary outcome data for only 187 of them, oral sodium bicarbonate did not improve physical function or kidney function, and adverse events were more frequent. (Source 19)

• Randomized trial, Moderate certainty.
• Size: 300 participants, primary outcome data for 187.
• Who: Adults aged 60 and over with estimated glomerular filtration rate under 30 mL/min/1.73 m2, not on dialysis, serum bicarbonate under 22 mmol/L, from 27 UK departments.
• How long: Up to 2 years, primary outcome at 12 months.
• Result: Short Physical Performance Battery adjusted treatment effect -0.4 (95% CI -0.9 to 0.1, p = 0.15); glomerular filtration rate 0.6 mL/min/1.73 m2 (95% CI -0.8 to 2.0, p = 0.39); adverse events 457 versus 400; higher costs and lower EQ-5D-3L quality of life with bicarbonate.
• Funding: Research Support, Non-U.S. Gov't per the record; a UK NHS-facing cost-effectiveness analysis was included.

Limit of this finding: Only 187 of the 300 randomised participants had primary outcome data, so the analysed trial was smaller than the headline number. The physical-function result is a null with the point estimate slightly favouring placebo (-0.4, 95% CI -0.9 to 0.1, p=0.15); that is a failure to show benefit, not a demonstration that bicarbonate and placebo are equivalent. And "adverse events were more frequent" is the authors' own wording resting on raw event counts, 457 against 400, with no denominator-adjusted rate and no statistical test printed in the abstract.

We found no significant treatment effect for the SPPB: bicarbonate arm 8.3 (SD 2.5) points, placebo arm 8.8 (SD 2.2) and adjusted treatment effect - 0.4 (95% CI - 0.9 to 0.1, p = 0.15).

In a 2026 double-blind randomised trial in 21 Danish hospitals, sodium bicarbonate given during in-hospital cardiac arrest did not increase return of spontaneous circulation. (Source 5)

• Randomized trial, High certainty.
• Size: 913 randomised, 779 in the primary analyses (372 bicarbonate, 407 placebo)
• Who: Adults with in-hospital cardiac arrest who had received at least one dose of epinephrine; median age 73 years, 64% male.
• How long: Enrolment February 2023 to February 2026, follow-up to 90 days.
• Result: Sustained return of spontaneous circulation 146/372 (39%) versus 150/407 (37%), risk ratio 1.05 (95% CI 0.88 to 1.24), P = .62, an absolute difference of about 2 percentage points; alive at 30 days 12% versus 9.1% (risk ratio 1.25, 95% CI 0.84 to 1.88); favourable neurologic outcome 8.1% versus 5.4% (risk ratio 1.39, 95% CI 0.82 to 2.34)
• Funding: not stated in the abstract.

Sustained return of spontaneous circulation occurred in 146 patients (39%) in the sodium bicarbonate group and 150 (37%) in the placebo group (risk ratio, 1.05 [95% CI, 0.88-1.24]; P = .62).

A meta-analysis of six observational studies in 18,406 adult cardiac arrest patients found no improvement in return of circulation or survival to discharge with sodium bicarbonate. (Source 20)

• Meta-analysis, Very low certainty.
• Size: 6 observational studies, 18,406 patients.
• Who: Non-traumatic adult cardiac arrest patients.
• How long: Databases searched from inception to December 2019.
• Result: Return of spontaneous circulation odds ratio 1.185 (95% CI 0.680 to 2.065); survival to discharge odds ratio 0.296 (95% CI 0.066 to 1.323). In a North American subgroup both rates were significantly lower with bicarbonate; these are observational data in which sicker patients are likelier to be given it.
• Funding: not stated.

There were no significant differences in the ROSC rate (odds ratio [OR] 1.185; 95% confidence interval [CI] 0.680-2.065) and survival-to-discharge rate (OR 0.296; 95% CI 0.066-1.323) between the SB and no-SB groups.

In the BICAR-ICU trial, sodium bicarbonate infusion for severe metabolic acidaemia did not improve the primary composite of death by day 28 and organ failure at day 7, and caused more alkalosis, high sodium and low calcium. (Source 21)

• Randomized trial, Moderate certainty.
• Size: 389 patients (194 control, 195 bicarbonate) across 26 French intensive care units.
• Who: Adults admitted within 48 h to intensive care with pH 7.20 or below, PaCO2 45 mm Hg or below and sodium bicarbonate concentration 20 mmol/L or below, plus organ failure or raised lactate.
• How long: Enrolment May 2015 to May 2017, outcomes to day 28.
• Result: Primary outcome 138/194 (71%) control versus 128/195 (66%) bicarbonate, absolute difference -5.5% (95% CI -15.2 to 4.2), p=0.24. In the prespecified acute kidney injury stratum with AKIN score 2 or 3, day 28 survival was 37% versus 54% (p=0.0283). Metabolic alkalosis, hypernatraemia and hypocalcaemia were more frequent with bicarbonate, with no life-threatening complications reported.
• Funding: French Ministry of Health and the Societe Francaise d'Anesthesie Reanimation.

Limit of this finding: The acute kidney injury result is a prespecified subgroup, not the trial's primary outcome, which was not improved. Note also an asymmetry inside the source itself: its Interpretation section says bicarbonate "decreased the primary composite outcome and day 28 mortality" in that stratum, while its Findings section reports only day-28 survival figures for the stratum (37% against 54%, p=0.0283) and gives no stratum-specific composite-outcome numbers.

The primary outcome occurred in 138 (71%) of 194 patients in the control group and 128 (66%) of 195 in the bicarbonate group (absolute difference estimate -5·5%, 95% CI -15·2 to 4·2; p=0·24).

A 2025 meta-analysis of blinded placebo-controlled running trials found the performance benefit of a single oral dose of sodium bicarbonate to be negligible and not statistically significant once gut-related dropouts and publication bias were allowed for. (Source 14)

• Meta-analysis, Moderate certainty.
• Size: 11 crossover studies, 126 participants.
• Who: Mostly male (84%) runners; typical dose 0.3 g/kg; performance tests 1 to 30 minutes, median 4 minutes.
• How long: Single doses; trials published through 31 December 2024.
• Result: Adjusted standardised mean difference 0.18 (95% CI -0.01 to 0.36), p = 0.06, I squared 0%, moderate certainty, which the authors' own scale classes as negligible. In the 8 male-only studies the effect was small and significant (SMD 0.40, 95% CI 0.18 to 0.63, p < 0.001)
• Funding: not stated.

After adjusting for GI-related study withdrawal and publication bias, the treatment effect of SB was negligible and not statistically significant (SMD = 0.18; 95% CI: -0.01, 0.36; p = 0.06; I2 = 0%; moderate certainty).

A Cochrane review of sodium bicarbonate for acute kidney injury found no randomised trial that met its criteria, so there is no randomised evidence either way for that use. (Source 22)

• Systematic review, Certainty not rated.
• Size: No studies included; four were identified by the search and none met the selection criteria.
• Who: Adults with acute kidney injury.
• How long: The 2012 MEDLINE abstract recorded here prints no search date. The currently published version of the same review (DOI 10.1002/14651858.CD009204.pub2, update dated June 2017, PMC6481784) states: "We searched the Cochrane Kidney and Transplant Specialised Register up to 12 May 2017"
• Result: No effect estimate exists. The review set out to assess all-cause mortality, need for renal replacement therapy, return to baseline kidney function and overall survival, and could pool nothing. Its authors call for well conducted randomised trials.
• Funding: not stated in the abstract.

Limit of this finding: An empty review is an absence of evidence, not evidence of absence: it shows that nobody had run a qualifying randomised trial by the review's search date, not that sodium bicarbonate does not work in acute kidney injury. On the date: the 2012 abstract this passage is taken from carries no search date, and the currently published version of the same review (same DOI, updated 2017) says its search ran up to 12 May 2017, through what is now called the Cochrane Kidney and Transplant Specialised Register rather than the Cochrane Renal Group's register. So the empty-review position is current to May 2017. The review also does not cover prevention, which it explicitly excluded, and acute kidney injury is a different setting from the chronic kidney disease trials reported above. No later version exists - CD009204.pub3 does not resolve.

We found no RCT evidence - supportive or otherwise - for the use of sodium bicarbonate for people with AKI. We concluded that there is an urgent need for well conducted RCTs in this area.

Where the evidence is mixed

A 2025 systematic review and meta-analysis found sodium bicarbonate raised serum bicarbonate in chronic kidney disease but did not significantly improve estimated glomerular filtration rate. (Source 1)

• Meta-analysis, Very low certainty.
• Size: 20 randomised and 2 non-randomised trials, 2,932 patients.
• Who: Patients with chronic kidney disease and metabolic acidosis.
• How long: Databases searched from inception to May 2024.
• Result: Serum bicarbonate mean difference 2.59 (95% CI 0.95 to 4.22, P = 0.02) with I squared 95%; eGFR mean difference 0.93 (95% CI -1.88 to 3.75, P = 0.52) with I squared 93%. Heterogeneity was extreme for both, which the review's own numbers show.
• Funding: not stated.

However, there was a non-significant increase in estimated glomerular filtration rate (eGFR) in patients on sodium bicarbonate therapy (MD: 0.93, 95%CI: -1.88-3.75; P = 0.52; I 2 = 93%).

A systematic review of oral bicarbonate regimens for urine alkalinisation before high-dose methotrexate found no significant differences against intravenous protocols except on two outcomes: urine pH falling below 7, and the incidence of diarrhoea, which was nearly three times the odds. (Source 23)

• Systematic review, Very low certainty.
• Size: 12 studies with 1,212 participants; 8 pooled for meta-analysis.
• Who: Patients receiving high-dose methotrexate.
• How long: Not stated; included randomised trials, cohorts, case reports and case series.
• Result: No significant differences in mean differences or odds ratios except for urine pH falling below 7 (mean difference 0.91, 95% CI 0.32 to 1.5, P < 0.05) and the incidence of diarrhoea (odds ratio 2.92, 95% CI 1.69 to 5.05, P < 0.05). The review does NOT say which arm had more episodes of urine pH below 7. Most included studies were non-randomised and were assessed with the ROBINS-I tool.
• Funding: not stated.

Limit of this finding: The review names the urine-pH-below-7 outcome and gives an effect size for it, but never states which arm had more - and a "mean difference" on an outcome phrased as an event is ambiguous. We could not resolve the direction: the article is not open access (Europe PMC records PMID 38233724 as closed) and both full-text routes we tried for PMC11087431 returned HTTP 404 and 500, so no table or forest plot was reachable. The direction has therefore been held rather than guessed. The diarrhoea result is unambiguous and is on the oral regimen.

No significant differences in mean differences (MDs) or odds ratio (OR) were found after the oral bicarbonate regimen, except for when urine pH fell to < 7 (MD: 0.91, 95% CI: 0.32, 1.5, P < 0.05) and the incidence of diarrhea (OR: 2.92, 95% CI: 1.69, 5.05, P < 0.05).

Where the research disagrees

Whether oral sodium bicarbonate slows chronic kidney disease or helps people with it feel or function better

• Authors of the 2009 bicarbonate supplementation trial, Two-year randomised trial in 134 adults against standard care, not placebo, and not blinded; outcome was creatinine clearance and progression. The relative risks the paper prints do not match the ratios of the percentages it prints, so the size of the effect is less certain than the figures suggest: fewer patients supplemented with bicarbonate developed ESRD (6.5 versus 33%; relative risk 0.13; 95% confidence interval 0.04 to 0.40; P < 0.001) (Source 3)
• Authors of the BiCARB trial, Pragmatic multicentre double-blind placebo-controlled randomised trial, 300 participants aged 60 and over, 27 UK departments, up to 2 years: Oral sodium bicarbonate did not improve physical function or renal function, increased adverse events and is unlikely to be cost-effective for use by the UK NHS for this patient group. (Source 12)

Whether sodium bicarbonate improves high-intensity exercise performance

• Authors of the cycling time-trial meta-analysis, Systematic review and meta-analysis of 18 randomised placebo-controlled cycling studies, no adjustment for publication bias reported: Overall, when compared to placebo, the NaHCO3 ingestion improved the Wmean (SMD: 0.42; 95% CI: 0.21-0.63; P = 0.001) and performance time (SMD: 0.22; 95% CI: 0.02-0.43; P = 0.03). (Source 6)
• Authors of the running meta-analysis, Systematic review and meta-analysis of 11 randomised double-blind placebo-controlled crossover trials with intent-to-treat adjustment for gut-related withdrawals and trim-and-fill correction for publication bias, GRADE moderate certainty: SB supplementation has a negligible benefit on continuous running performance in a mixed-sex population, the ergogenic effect may be more pronounced in males, GI symptoms are common, and some users may not tolerate supplementation. (Source 24)

How much

• Reference intake: No dietary reference intake applies: this is a medicine. The over-the-counter amount is set by the Drug Facts directions, and the kidney, intensive-care and urinary uses are set by a prescriber. The US label published 25 September 2026 for a 650 mg sodium bicarbonate antacid tablet directs adults under 60 to 1 to 4 tablets every 4 hours, and adults 60 and over to 1 to 2 tablets every 4 hours. (Source 7)
• Upper limit: The same US label caps its own use at 24 tablets in 24 hours for adults up to 60 years of age, 12 tablets for adults 60 and over, and not more than 2 weeks at maximum dosage without a physician's advice. This is a labelling position dated 25 September 2026. (Source 7)
• Studied: The BiCARB trial gave participants oral sodium bicarbonate up to 3 g a day, or matching placebo, for up to 2 years. (Source 25)
• Studied: BICAR-ICU infused 4.2% intravenous sodium bicarbonate to keep arterial pH above 7.30, with each infusion 125 to 250 mL over 30 min and a maximum of 1000 mL in 24 hours. (Source 21)
• Studied: The 2026 cardiac arrest trial gave sodium bicarbonate up to 100 mmol or placebo intravenously during the arrest. (Source 26)
• Studied: The exercise trials typically used a single oral dose of 0.3 g/kg, and the blood pressure and gut-distress crossover trial used 300 mg/kg. (Source 14)
• Studied: The urinary alkalinisation pilot gave 33 women 2 x 4 g a day for 4 weeks. (Source 4)

A common belief, and what the research shows

The belief: Baking soda is harmless because it is a kitchen ingredient, and it is standard treatment during a cardiac arrest.

What the research shows: Neither holds up. In the largest randomised trial of sodium bicarbonate in in-hospital cardiac arrest, return of spontaneous circulation occurred in 39% of the bicarbonate group and 37% of the placebo group (risk ratio 1.05, 95% CI 0.88 to 1.24, P = .62), and the authors concluded that the findings do not support routine administration. On harms, a single ergogenic dose caused diarrhoea in 10 of 11 trained volunteers, and a published series of 36 spontaneous gastric ruptures, 10 of them associated with sodium bicarbonate ingestion, carried a mortality of 41.6%.

Questions and answers

What is it?

Sodium bicarbonate is baking soda, a simple alkaline sodium salt. Sold as a medicine it comes as 650 mg antacid tablets, as powder for alkalinising urine, and as a solution for injection in hospitals. A US Drug Facts label gives the tablet strength and its purpose as antacid. (Source 7)

What does it do in the body?

It is an alkali. Absorbed into the blood it raises serum bicarbonate, which is how it corrects the acidosis of advanced kidney disease; a 2025 meta-analysis of 22 studies in 2,932 patients confirmed that rise, with a mean difference of 2.59, though the trials disagreed with each other a great deal. Its acid-neutralising action in the stomach, and its effect on urine pH, are set out under those uses on this page, each with its own source. (Source 1)

Is it good or bad for you?

It depends entirely on the use, and for several uses the honest answer is that it does not help. It reliably changes blood and urine chemistry. But in the blinded BiCARB trial in older people with advanced kidney disease it did not improve physical function or kidney function, adverse events were more frequent, and the authors judged it unlikely to be cost-effective. Its clearest benefits are chemical, not clinical. (Source 12)

How do you get more of it?

The antacid tablets are sold over the counter and the label sets the amount and the ceiling; the kidney, urinary and hospital uses are prescribed. The US label, as a position dated 25 September 2026, states the ceiling for its own use. None of this is advice for any individual. (Source 7)

If it is harmful, what reduces it?

The remedy for too much is stopping it and letting the kidneys clear the load; the specific harms reported are metabolic alkalosis and high blood sodium, which were more frequent on bicarbonate in both the BICAR-ICU trial and the 2026 cardiac arrest trial. In BICAR-ICU these were reported with no life-threatening complications. (Source 21)

Why might someone be low in it or missing it?

Does not apply in the way it would for a nutrient. Sodium bicarbonate as a medicine is a manufactured salt, not something a person can be deficient in. What can be low is the body's own blood bicarbonate, which falls in advanced kidney disease; the BiCARB trial recruited people with a serum bicarbonate concentration below 22 mmol/L and an estimated glomerular filtration rate under 30 mL/min/1.73 m2 for exactly that reason. (Source 25)

Which whole foods contain it or feed it?

Baking soda is a kitchen ingredient used as a raising agent, so it does reach people through cooked foods, but we found no study in this search that measured dietary sodium bicarbonate or compared a food source with the medicinal dose. What the literature does quantify is the sodium that comes with it: a labelled 650 mg tablet carries sodium 178 mg. (Source 7)

We searched: Europe PMC searches for sodium bicarbonate with antacid, exercise performance, blood pressure, sodium load, gastric rupture, urinary alkalinisation and chronic kidney disease; no study of dietary or whole-food sodium bicarbonate was found.

What happens if you do not have it?

In the 2026 randomised cardiac arrest trial, going without made no measurable difference: return of spontaneous circulation occurred in 150 of 407 placebo patients (37%) against 146 of 372 on sodium bicarbonate (39%), a risk ratio of 1.05 with a confidence interval from 0.88 to 1.24 and P = .62. Survival at 30 days and favourable neurologic outcome were also not significantly different. For the drug's other uses, what happens without it is set out under those uses, each with its own source. (Source 5)

How can you test for it?

Two tests matter, and both measure the effect rather than the drug. A blood test for serum bicarbonate shows whether acidosis is corrected, and trials use it as their entry criterion and their chemical endpoint. A urine pH strip or laboratory measurement shows whether urine has been alkalinised; in the 33-woman pilot, oral sodium bicarbonate raised urine pH from 5.31 to 7.2 while blood pH did not change significantly, so a urine result does not tell you about blood chemistry or the reverse. (Source 18)

References

1. World journal of nephrology. Safety and efficacy of sodium bicarbonate for treating metabolic acidosis in chronic kidney disease: A systematic review and meta-analysis — abstract, Results section. 2025. PMID 40134641, DOI 10.5527/wjn.v14.i1.101078. Read the source
2. American journal of therapeutics. Self-Directed Treatment of Intermittent Heartburn: A Randomized, Multicenter, Double-Blind, Placebo-Controlled Evaluation of Antacid and Low Doses of an H(2)-Receptor Antagonist (Famotidine) — abstract, RESULTS section of the abstract. 1995. PMID 11850668, DOI 10.1097/00045391-199505000-00003. Read the source
3. Journal of the American Society of Nephrology : JASN. Bicarbonate supplementation slows progression of CKD and improves nutritional status — abstract. 2009. PMID 19608703, DOI 10.1681/asn.2008111205. Read the source
4. International urogynecology journal. Effects of urine alkalinization with sodium bicarbonate orally on lower urinary tract symptoms in female patients: a pilot study — abstract, Methods section. 2018. PMID 28975365, DOI 10.1007/s00192-017-3492-3. Read the source
5. JAMA. Sodium Bicarbonate for In-Hospital Cardiac Arrest: A Randomized Clinical Trial — abstract, Results section. 2026. PMID 42273960, DOI 10.1001/jama.2026.10628. Read the source
6. European journal of sport science. Acute effects of sodium bicarbonate ingestion on cycling time-trial performance: A systematic review and meta-analysis of randomized controlled trials — abstract, abstract body, up to the end of the concluding sentence. 2023. PMID 35633035, DOI 10.1080/17461391.2022.2071171. Read the source
7. DailyMed (US National Library of Medicine), SPL from Rugby Laboratories (Indianapolis, IN, USA). Sodium Bicarbonate Antacid 10 gr (650 mg) tablets - US OTC Drug Facts label, complete content from "Drug Facts" through "Other information"; each SPL section title, and list is on its own line, so "Other information" is followed by its two list items, the sodium content and the storage line. Sep 25, 2026. Read the source
8. Nutrition research (New York, N.Y.). Acute sodium bicarbonate loading has negligible effects on resting and exercise blood pressure but causes gastrointestinal distress — abstract, first part of the abstract (rationale, design and blood pressure results). 2013. PMID 23746564, DOI 10.1016/j.nutres.2013.04.009. Read the source
9. Journal of eating disorders. Binge-eating and sodium bicarbonate: a potent combination for gastric rupture in adults-two case reports and a review of literature — abstract, Methods section. 2022. PMID 36348449, DOI 10.1186/s40337-022-00677-9. Read the source
10. Daru : journal of Faculty of Pharmacy, Tehran University of Medical Sciences. Oral regimen for high dose methotrexate urine alkalinization: a systematic review and meta-analysis — abstract, Objective section. 2024. PMID 38233724, DOI 10.1007/s40199-023-00499-3. Read the source
11. Journal of eating disorders. Binge-eating and sodium bicarbonate: a potent combination for gastric rupture in adults-two case reports and a review of literature — abstract, Results section. 2022. PMID 36348449, DOI 10.1186/s40337-022-00677-9. Read the source
12. BMC medicine. Clinical and cost-effectiveness of oral sodium bicarbonate therapy for older patients with chronic kidney disease and low-grade acidosis (BiCARB): a pragmatic randomised, double-blind, placebo-controlled trial — abstract, Conclusions section. 2020. PMID 32268897, DOI 10.1186/s12916-020-01542-9. Read the source
13. JAMA. Sodium Bicarbonate for In-Hospital Cardiac Arrest: A Randomized Clinical Trial — abstract, Conclusions and relevance section. 2026. PMID 42273960, DOI 10.1001/jama.2026.10628. Read the source
14. Journal of the International Society of Sports Nutrition. Negligible benefit of oral single-dose sodium bicarbonate on continuous running performance: systematic review with meta-analysis of randomized, double-blind, placebo-controlled trials — abstract, Results section. 2025. PMID 41416636, DOI 10.1080/15502783.2025.2538606. Read the source
15. Nutrition research (New York, N.Y.). Acute sodium bicarbonate loading has negligible effects on resting and exercise blood pressure but causes gastrointestinal distress — abstract, latter part of the abstract (gastrointestinal results and conclusion). 2013. PMID 23746564, DOI 10.1016/j.nutres.2013.04.009. Read the source
16. DailyMed (US National Library of Medicine), SPL from Rugby Laboratories (Indianapolis, IN, USA). Sodium Bicarbonate Antacid 10 gr (650 mg) tablets - US OTC label, the whole SPL UNCLASSIFIED SECTION (LOINC 42229-5), from "TAMPER EVIDENT:" through the California Proposition 65 warning and the closing revision line; this section sits outside the Drug Facts panel. Sep 25, 2026. Read the source
17. Clinical journal of the American Society of Nephrology : CJASN. Effect of Sodium Bicarbonate on Systolic Blood Pressure in CKD: A Systematic Review and Meta-Analysis — abstract, Results section. 2023. PMID 36758154, DOI 10.2215/cjn.0000000000000119. Read the source
18. International urogynecology journal. Effects of urine alkalinization with sodium bicarbonate orally on lower urinary tract symptoms in female patients: a pilot study — abstract, Results section. 2018. PMID 28975365, DOI 10.1007/s00192-017-3492-3. Read the source
19. BMC medicine. Clinical and cost-effectiveness of oral sodium bicarbonate therapy for older patients with chronic kidney disease and low-grade acidosis (BiCARB): a pragmatic randomised, double-blind, placebo-controlled trial — abstract, Results section. 2020. PMID 32268897, DOI 10.1186/s12916-020-01542-9. Read the source
20. The Journal of emergency medicine. Effectiveness of Sodium Bicarbonate Administration on Mortality in Cardiac Arrest Patients: A Systematic Review and Meta-analysis — abstract, Results section. 2020. PMID 32978028, DOI 10.1016/j.jemermed.2020.08.012. Read the source
21. Lancet (London, England). Sodium bicarbonate therapy for patients with severe metabolic acidaemia in the intensive care unit (BICAR-ICU): a multicentre, open-label, randomised controlled, phase 3 trial — abstract, Findings section. 2018. PMID 29910040, DOI 10.1016/s0140-6736(18)31080-8. Read the source
22. The Cochrane database of systematic reviews. Sodium bicarbonate supplements for treating acute kidney injury (Cochrane review CD009204.pub2) — abstract, the Objectives, Selection criteria, Main results and Authors' conclusions sections, each under its own heading. 2012. PMID 22696382, DOI 10.1002/14651858.CD009204.pub2. Read the source
23. Daru : journal of Faculty of Pharmacy, Tehran University of Medical Sciences. Oral regimen for high dose methotrexate urine alkalinization: a systematic review and meta-analysis — abstract, Results section. 2024. PMID 38233724, DOI 10.1007/s40199-023-00499-3. Read the source
24. Journal of the International Society of Sports Nutrition. Negligible benefit of oral single-dose sodium bicarbonate on continuous running performance: systematic review with meta-analysis of randomized, double-blind, placebo-controlled trials — abstract, Conclusions section. 2025. PMID 41416636, DOI 10.1080/15502783.2025.2538606. Read the source
25. BMC medicine. Clinical and cost-effectiveness of oral sodium bicarbonate therapy for older patients with chronic kidney disease and low-grade acidosis (BiCARB): a pragmatic randomised, double-blind, placebo-controlled trial — abstract, Methods section. 2020. PMID 32268897, DOI 10.1186/s12916-020-01542-9. Read the source
26. JAMA. Sodium Bicarbonate for In-Hospital Cardiac Arrest: A Randomized Clinical Trial — abstract, two consecutive labelled sections reproduced one after the other, each with its own MEDLINE/Europe PMC section label on its own line: "Design, setting, and participants" then "Intervention". The labels are reproduced in the down-cased form the fetched Europe PMC copy prints; JAMA's own JATS abstract prints them in title case ("Design, Setting, and Participants"). 2026. PMID 42273960, DOI 10.1001/jama.2026.10628. Read the source