Publication

• Title: Sodium Bicarbonate for In-Hospital Cardiac Arrest: A Randomized Clinical Trial
• Acronym: BIHCA
• Year: 2026
• Journal published in: JAMA
• Citation: Granfeldt A, Kirkegaard BL, Vallentin MF, Holmberg MJ, Stankovic N, Lind PC, et al; Bicarbonate for In-Hospital Cardiac Arrest (BIHCA) Investigators. Sodium bicarbonate for in-hospital cardiac arrest: a randomized clinical trial. JAMA. Published online June 11, 2026.

Context & Rationale

• Background

  • In-hospital cardiac arrest has poor outcomes despite rapid recognition and hospital-based advanced life support.
  • In the BIHCA population, all participants had received at least one dose of adrenaline during resuscitation, identifying a higher-risk subgroup than all-comer in-hospital cardiac arrest.
  • Severe acidaemia develops during cardiac arrest and may plausibly impair myocardial contractility, blunt catecholamine responsiveness, and worsen post-ischaemic injury.
  • Sodium bicarbonate can increase pH and buffer metabolic acidosis, providing a biologically intuitive rationale for use during prolonged resuscitation.
  • The countervailing physiological concerns include sodium loading, hyperosmolarity, paradoxical intracellular acidosis from carbon dioxide generation, alkalosis, hypernatraemia, hypocalcaemia, and distraction from interventions with proven benefit.
  • Routine sodium bicarbonate use has persisted despite modern guidance discouraging routine use; the accompanying editorial highlighted use in approximately 40% to 50% of in-hospital cardiac arrests and 19% of out-of-hospital cardiac arrests. ¹
  • Before BIHCA, there was no contemporary, adequately powered randomised trial testing sodium bicarbonate during adult in-hospital cardiac arrest.
• Research Question/Hypothesis

  • The trial tested whether routine intravenous sodium bicarbonate during adult in-hospital cardiac arrest, after adrenaline had been administered, would increase sustained return of spontaneous circulation compared with placebo.
  • The secondary hypothesis was that improved intra-arrest acid–base status would translate into higher 30-day survival and higher 30-day survival with favourable neurological outcome.
  • The trial therefore addressed a specific resuscitation pharmacology question: routine buffer therapy during in-hospital cardiac arrest, not targeted bicarbonate for hyperkalaemia, sodium-channel blocker poisoning, tricyclic antidepressant poisoning, or another clear clinical indication.
• Why This Matters

  • Sodium bicarbonate is cheap, familiar, and physiologically appealing, but widespread use without robust outcome benefit can still add complexity, treatment burden, and biochemical harm.
  • Routine use during cardiac arrest is a high-volume intervention where even a small effect on return of spontaneous circulation or survival would matter.
  • Conversely, a well-conducted trial showing no clinically important effect would support deimplementation of routine use and sharpen attention on reversible causes and guideline-based resuscitation.
  • BIHCA is important because it supplies randomised in-hospital evidence for a practice that was previously guided largely by physiology, older out-of-hospital trials, observational studies, and clinician preference.

Design & Methods

• Research Question: In adults with in-hospital cardiac arrest who had received at least one dose of adrenaline, does routine sodium bicarbonate, compared with placebo, increase sustained return of spontaneous circulation?
• Study Type: Investigator-initiated, multicentre, randomised, parallel-group, double-blind, placebo-controlled, superiority trial conducted in 21 Danish hospitals.
• Population:
  • Adults aged ≥18 years with in-hospital cardiac arrest.
  • Participants had received at least one dose of adrenaline during the cardiac arrest.
  • Enrolment occurred during active resuscitation after the first adrenaline dose.
  • Key exclusions were clearly documented do-not-resuscitate order before arrest, previous enrolment, invasive mechanical circulatory support at the time of arrest, known or suspected pregnancy, known objection to trial participation, or a clinical indication for sodium bicarbonate administration.
  • Of 2913 screened in-hospital cardiac arrests, 1855 had received at least one dose of adrenaline, 913 were randomised, and 779 were eligible for the primary analyses.
• Intervention:
  • Intravenous sodium bicarbonate 50 mL of 8.4% solution, equivalent to 50 mmol, given as soon as possible after the first dose of adrenaline.
  • If the patient remained in cardiac arrest, one additional 50 mmol dose was given after the next adrenaline dose.
  • The maximum trial dose was therefore 100 mmol sodium bicarbonate.
  • The intervention was a fixed-dose, immediate intra-arrest strategy rather than weight-based or blood-gas-guided therapy.
• Comparison:
  • Matching placebo using 50 mL of 0.9% sodium chloride from identical vials.
  • If the patient remained in cardiac arrest, one additional placebo dose was given after the next adrenaline dose.
  • All other resuscitation care was delivered by the clinical cardiac arrest team according to usual advanced life support practice.
  • Sodium bicarbonate outside the trial protocol was permitted when the clinical team judged it necessary, but occurred at similar frequency in both groups.
• Blinding:
  • Patients, clinicians, investigators, and outcome assessors were blinded.
  • Study drug was supplied in blinded kits with identical vials, retrieved by a dedicated cardiac arrest team member.
  • Blinding was a major methodological strength because it minimised differential resuscitation decisions, outcome assessment bias, and clinician expectancy effects.
• Statistics:
  • A total of 778 patients were required to detect an increase in sustained return of spontaneous circulation from 33% with placebo to 43% with sodium bicarbonate, corresponding to RR 1.30, with 80% power at a two-sided alpha of 0.05.
  • The primary analysis was a modified intention-to-treat analysis including patients who received the first dose of trial drug, met all inclusion criteria, had no exclusion criteria at randomisation, and had available outcome data.
  • Effect estimates were adjusted for prespecified strong prognostic factors: age, witnessed status, and initial rhythm.
  • Primary and key secondary outcomes were tested in hierarchical order; because the primary outcome was not statistically significant, p values were not calculated for key secondary outcomes.
  • An independent data monitoring committee reviewed safety after approximately 200 and 400 enrolled patients.
  • The trial had no prespecified stopping criteria for harm, futility, or efficacy.
• Follow-Up Period:
  • The primary outcome was assessed during the index resuscitation episode.
  • Key secondary outcomes were assessed at 30 days.
  • Additional survival, neurological outcome, and quality-of-life outcomes were reported at 90 days.
  • Collection of 180-day and 1-year outcomes was ongoing at the time of the main JAMA report.

Key Results

This trial continued to completion. It reached its planned analysable sample size; it was not stopped early for harm, futility, or efficacy.

• Sodium bicarbonate changed early post-arrest biochemistry but did not increase sustained return of spontaneous circulation.
• Survival and favourable neurological outcome were numerically higher with sodium bicarbonate, but confidence intervals were wide and these outcomes were not formally tested after the primary endpoint was not significant.
• The principal measurable harm was biochemical: alkalosis and hypernatraemia were more common after sodium bicarbonate.
• Primary outcome subgroup analyses did not identify a group with clear benefit: non-shockable rhythm RR 1.04 (95% CI 0.85 to 1.26), shockable rhythm RR 1.23 (95% CI 0.80 to 1.88), known pre-arrest metabolic acidosis RR 1.00 (95% CI 0.65 to 1.54), time to trial drug <8 minutes RR 1.21 (95% CI 0.95 to 1.56), and witnessed arrest RR 1.09 (95% CI 0.91 to 1.31).

Internal Validity

• Randomisation and Allocation:
  • Randomisation used a computer-generated 1:1 allocation sequence.
  • Allocation concealment was strong because blinded kits contained identically labelled vials.
  • The published article reported block randomisation in fixed blocks of 6 without stratification.
  • The supplemental appendix notes a protocol discrepancy: the protocol had stated random block sizes of 2, 4, or 6 and stratification by site, but for practical reasons randomisation was not stratified by site and used blocks of 6. ²
  • This discrepancy is unlikely to materially bias the result because allocation remained concealed and the trial was double-blind.
• Dropout and Post-randomisation Exclusions:
  • Of 913 randomised patients, 779 were included in the primary analyses: 372 assigned to sodium bicarbonate and 407 assigned to placebo.
  • Three patients had lost randomisation numbers.
  • In the sodium bicarbonate group, 57 randomised patients did not receive trial drug: 24 achieved ROSC before administration, 13 had resuscitation terminated, and 20 had other reasons.
  • In the placebo group, 55 randomised patients did not receive trial drug: 27 achieved ROSC before administration, 7 had resuscitation terminated, and 21 had other reasons.
  • Thirteen patients in the sodium bicarbonate group and 5 in the placebo group received trial drug but were excluded because they did not meet inclusion criteria or met exclusion criteria that were not known to the clinical team at randomisation.
  • This modified intention-to-treat strategy is a legitimate point of critique because it excludes randomised patients after allocation.
  • The editorial judged this less problematic than a conventional per-protocol analysis because most excluded patients did not receive blinded trial drug, making a biological drug effect implausible in those cases. ¹
• Performance and Detection Bias:
  • The double-blind design markedly reduced performance bias during resuscitation.
  • The primary outcome, sustained ROSC for at least 20 minutes, was objective and temporally close to the intervention.
  • Longer-term outcomes are more vulnerable to post-randomisation care pathways, prognostication, coronary intervention decisions, intensive care limitations, and withdrawal of life-sustaining treatment.
  • The trial reported hospital disposition and causes of death, which helps interpret post-randomisation treatment decisions.
• Protocol Adherence:
  • Protocol delivery was good but not perfect.
  • Two trial drug doses were given to 256/372 (69%) in the sodium bicarbonate group and 289/407 (71%) in the placebo group.
  • One dose was given to 116/372 (31%) and 118/407 (29%), respectively.
  • Among patients still in cardiac arrest after two adrenaline doses, the second trial drug dose was not administered to 66/372 (18%) in the sodium bicarbonate group and 63/407 (15%) in the placebo group.
  • Sodium bicarbonate outside the protocol was administered during cardiac arrest to 23/372 (6.2%) in the sodium bicarbonate group and 27/407 (6.6%) in the placebo group. ²
• Baseline Characteristics:
  • Groups were broadly comparable.
  • Median age was 74 years in the sodium bicarbonate group and 73 years in the placebo group.
  • Male sex was 65% versus 64%.
  • Initial rhythms were mostly non-shockable: pulseless electrical activity 53% versus 51%, and asystole 33% versus 36%.
  • Known metabolic acidosis before cardiac arrest was present in 53/372 (14%) versus 53/407 (13%).
  • There were small imbalances that favour neither group cleanly: monitored arrest was 47% versus 39%, witnessed arrest 76% versus 71%, and coronary artery disease 30% versus 25%.
  • Adjustment for age, witnessed status, and initial rhythm was prespecified and appropriate.
  • The cohort was sufficiently ill to detect a plausible resuscitation effect: 88% had non-shockable rhythms and 30-day survival was only 9.1% to 12%.
• Heterogeneity:
  • The trial population was clinically heterogeneous by cause of arrest: pulmonary 35% versus 30%, cardiac 29% versus 29%, hypotension or hypovolaemia 11% versus 10%, and unknown 20% versus 27%.
  • This heterogeneity strengthens pragmatic relevance but dilutes any effect limited to specific acid–base or toxicological phenotypes.
  • The trial did not require documentation of severe intra-arrest acidaemia before randomisation.
  • Known metabolic acidosis before cardiac arrest was defined using pre-arrest pH and base excess and was present in only 106 analysed patients.
  • Subgroup analyses were therefore underpowered for the most biologically appealing subgroup.
• Timing:
  • Drug delivery was early for a cardiac arrest pharmacology trial.
  • Median time from arrest to adrenaline was 6 minutes in the sodium bicarbonate group and 5 minutes in the placebo group.
  • Median time from arrest to trial drug was 8 minutes in both groups.
  • This timing was appropriate for testing the hypothesis that early buffering might improve ROSC.
  • The absence of a clear effect despite early administration weakens the argument that previous out-of-hospital trials failed only because drug delivery was too late.
• Dose:
  • The dose was fixed at 50 mmol, repeated once if the patient remained in cardiac arrest, for a maximum of 100 mmol.
  • This was feasible during resuscitation and aligned with pragmatic emergency drug delivery.
  • The dose achieved biochemical separation: first post-ROSC pH was 7.11 versus 7.03 and standard bicarbonate was 17.7 mmol/L versus 13.1 mmol/L.
  • The dose did not normalise acid–base status; median pH after ROSC remained severely acidotic in the sodium bicarbonate group.
  • Whether higher, repeated, or weight-based dosing would change outcomes remains unanswered, but higher doses would plausibly increase alkalosis, sodium load, and carbon dioxide generation.
• Separation of the Variable of Interest:
  • Trial drug exposure separated the groups well: sodium bicarbonate arm received up to 100 mmol bicarbonate, placebo arm received matching sodium chloride.
  • Post-ROSC standard bicarbonate separated by 3.1 mmol/L (95% CI 1.8 to 4.4).
  • Post-ROSC sodium separated by 3.2 mmol/L (95% CI 1.5 to 4.9).
  • Post-ROSC alkalosis was 35% versus 20%.
  • Post-ROSC hypernatraemia was 42% versus 29%.
  • These laboratory differences confirm that the intervention was biologically active, even though clinical outcomes did not improve.
• Key Delivery Aspects:
  • The intervention was delivered by existing cardiac arrest teams during real resuscitations, increasing clinical relevance.
  • The requirement for a dedicated team member to retrieve and prepare the blinded kit may limit implementation in smaller hospitals, but protected trial integrity.
  • The trial tested routine administration after adrenaline, not precision use guided by potassium, pH, base excess, toxidrome, or duration of arrest.
• Crossover:
  • Clinically directed bicarbonate outside the protocol occurred at nearly identical rates during cardiac arrest: 6.2% versus 6.6%.
  • Among those surviving at least 24 hours, sodium bicarbonate during the first 24 hours was given to 29/95 (31%) in the sodium bicarbonate group and 25/90 (28%) in the placebo group.
  • This degree of crossover is unlikely to explain the primary outcome.
• Adjunctive Therapy Use:
  • Intra-arrest co-interventions were balanced.
  • Median adrenaline doses were 3 (IQR 2 to 5) versus 3 (IQR 2 to 4).
  • Amiodarone was used in 54/372 (15%) versus 64/407 (16%).
  • Calcium was used in 37/372 (9.9%) versus 30/407 (7.4%).
  • Defibrillation occurred in 102/372 (27%) versus 110/407 (27%).
  • Intubation during arrest occurred in 262/372 (70%) versus 283/407 (70%).
  • Extracorporeal cardiopulmonary resuscitation occurred in 5/372 (1.3%) versus 11/407 (2.7%).
  • Post-arrest interventions among 24-hour survivors showed some imbalances, including percutaneous coronary intervention 4/95 (4%) versus 13/90 (14%) and VA-ECMO 6/95 (6%) versus 9/90 (10%); these post-randomisation differences matter more for long-term outcomes than for ROSC.
• Outcome Assessment:
  • Sustained ROSC was clearly defined as signs of circulation without chest compressions for at least 20 minutes.
  • Survival was objective.
  • Neurological outcome used the modified Rankin Scale, with favourable outcome defined as 0 to 3.
  • Quality of life used EQ-5D-5L among survivors, but survivor numbers were small.
  • Organ dysfunction and laboratory endpoints were objective.
• Statistical Rigor:
  • The sample-size target was met.
  • The primary outcome analysis was adjusted for strong prespecified prognostic factors.
  • The hierarchical testing approach protected against overinterpreting secondary outcomes.
  • A post hoc analysis including all randomised patients who received trial intervention found an unadjusted RR for ROSC of 1.04 (95% CI 0.88 to 1.24), consistent with the primary analysis. ²
  • The modified intention-to-treat analysis is transparent but remains a methodological limitation for purists who prefer all randomised patients to be retained in the primary analysis.
• Conclusion on Internal Validity:
  • Overall, internal validity is strong for the primary question of average treatment effect on sustained ROSC.
  • The main caveats are post-randomisation exclusions, the modified intention-to-treat population, limited power for survival and neurological outcomes, and unavoidable post-randomisation confounding for longer-term outcomes.

External Validity

• Population Representativeness:
  • The trial is highly applicable to adult in-hospital cardiac arrest patients who require adrenaline during resuscitation in high-resource hospital systems.
  • The trial population was older, with median age 73 years overall.
  • Most arrests occurred on hospital wards: 64% in the trial population.
  • Only 8% occurred in intensive care units, whereas US registry populations include a much larger intensive care component.
  • Most patients had non-shockable rhythms: 88% in the trial population.
  • The epinephrine requirement made this a selected, higher-risk population with worse outcomes than all-comer in-hospital cardiac arrest.
• Important Exclusions:
  • Patients with a clinical indication for sodium bicarbonate were excluded.
  • The findings should not be extrapolated to hyperkalaemic arrest, sodium-channel blocker poisoning, tricyclic antidepressant overdose, or other situations where bicarbonate has a specific mechanistic role.
  • Patients already receiving invasive mechanical circulatory support at the time of arrest were excluded.
  • Pregnant patients were excluded.
  • Children were excluded.
  • Out-of-hospital cardiac arrests were excluded.
  • Patients who achieved ROSC before adrenaline or did not receive adrenaline were not represented.
• Applicability:
  • The findings translate well to hospitals with organised cardiac arrest teams, early vascular access, rapid adrenaline administration, and rapid drug delivery.
  • Applicability may be lower in systems where in-hospital resuscitation drugs are delayed, arrest aetiologies differ, or resources for post-arrest care differ substantially.
  • The trial informs routine empiric sodium bicarbonate use during in-hospital cardiac arrest, not targeted correction of documented severe metabolic acidosis outside the arrest algorithm.
  • The trial does not settle buffer therapy in prolonged out-of-hospital cardiac arrest, paediatric arrest, toxicological arrest, renal failure with hyperkalaemia, or arrests managed with extracorporeal CPR protocols.
• Conclusion on External Validity:
  • External validity is high for adult in-hospital cardiac arrest requiring adrenaline in well-resourced systems similar to Denmark.
  • Generalisability is limited for special-circumstance bicarbonate indications, out-of-hospital cardiac arrest, paediatric arrest, and settings with substantially different resuscitation and post-arrest systems.

Strengths & Limitations

• Strengths:
  • First large randomised trial of sodium bicarbonate during adult in-hospital cardiac arrest.
  • Investigator-initiated national Danish trial across 21 hospitals.
  • Double-blind, placebo-controlled design with concealed allocation.
  • Early drug delivery: median 8 minutes from cardiac arrest to trial drug.
  • Clinically relevant primary endpoint proximate to the intervention.
  • High-quality ascertainment of survival and neurological outcomes.
  • Biochemical separation confirmed that the intervention was delivered and pharmacologically active.
  • Detailed reporting of intra-arrest therapies, post-arrest interventions, organ dysfunction, and adverse laboratory events.
• Limitations:
  • Powered for sustained ROSC, not for survival, favourable neurological outcome, or quality of life.
  • Modified intention-to-treat analysis excluded some randomised patients after allocation.
  • Large number of non-randomised patients with epinephrine-treated in-hospital cardiac arrest, mostly for logistical reasons, limits population representativeness.
  • Not stratified by site, despite earlier protocol language suggesting site stratification.
  • No requirement for intra-arrest or pre-randomisation blood gas confirmation of severe acidaemia.
  • Only a small subgroup had known metabolic acidosis before arrest.
  • Fixed dosing may not have been optimal for all body sizes or degrees of acidosis.
  • Conducted exclusively in Denmark.
  • Does not answer bicarbonate use for special circumstances such as hyperkalaemia or sodium-channel blocker poisoning.

Interpretation & Why It Matters

• Clinical practice

  BIHCA does not support routine sodium bicarbonate administration after adrenaline during adult in-hospital cardiac arrest.
• Mechanism

  Sodium bicarbonate produced the expected biochemical effect, with higher post-ROSC pH and standard bicarbonate, but this did not translate into higher sustained ROSC.
• Guideline alignment

  The trial strengthens the existing recommendation against routine bicarbonate during cardiac arrest while preserving the distinction between routine empiric use and selected special-circumstance use.
• Bedside implication

  For the average adult in-hospital arrest requiring adrenaline, bicarbonate should not be added reflexively because the patient is acidotic; clinicians should reserve it for specific indications.
• Why it changes thinking

  The trial separates physiological correction from patient benefit: improving blood gas values was not enough to improve ROSC, organ dysfunction, survival, or neurological outcome.

Controversies & Subsequent Evidence

• ROSC as the primary endpoint:
  • The most patient-centred outcomes are survival, neurological function, and quality of life.
  • BIHCA was powered for sustained ROSC because this outcome is close in time to the drug intervention and less affected by post-arrest care, prognostication, withdrawal of life support, coronary intervention, and intensive care capacity.
  • The accompanying editorial argued that the primary endpoint was appropriate for isolating the effect of an intra-arrest drug, while emphasising the need to report detailed downstream outcomes in resuscitation trials. ¹
  • The trade-off is that BIHCA cannot exclude smaller effects on survival or neurological outcome.
• Modified intention-to-treat analysis:
  • The primary analysis excluded randomised patients who did not receive trial drug and those later found to be ineligible.
  • This improves biological interpretability and precision for a drug administered during a rapidly changing resuscitation attempt.
  • It also means the primary analysis is not a pure intention-to-treat analysis.
  • The supplemental post hoc analysis including all randomised patients who received trial intervention was consistent with the primary result: unadjusted RR for ROSC 1.04; 95% CI 0.88 to 1.24. ²
• Routine use versus selected use:
  • BIHCA tested empiric routine sodium bicarbonate after adrenaline.
  • It did not test bicarbonate targeted to confirmed hyperkalaemia, sodium-channel blocker poisoning, tricyclic antidepressant overdose, severe pre-existing metabolic acidosis, or other special circumstances.
  • The 2025 ILCOR Advanced Life Support treatment recommendations suggested against buffering agents such as sodium bicarbonate for in-hospital cardiac arrest unless a special circumstance is present, while recognising that pre-BIHCA RCT data for in-hospital arrest did not exist. ³
  • The 2025 American Heart Association and European Resuscitation Council adult advanced life support guidelines were aligned with avoiding routine use before BIHCA was published. ⁴⁵
• Acidosis-targeted treatment remains unresolved:
  • The most biologically plausible beneficiaries are patients with severe metabolic acidosis, hyperkalaemia, or toxin-mediated sodium-channel blockade.
  • BIHCA did not require intra-arrest blood gas assessment before enrolment.
  • Known metabolic acidosis before cardiac arrest was present in only 53 patients in each group, with identical ROSC in that subgroup: 23/53 (43%) versus 23/53 (43%); RR 1.00; 95% CI 0.65 to 1.54.
  • The absence of a signal in this subgroup is informative but underpowered.
• Out-of-hospital extrapolation:
  • Out-of-hospital cardiac arrest differs by aetiology, time to drug, low-flow duration, post-arrest selection, and severity of acidaemia.
  • BIHCA delivered trial drug at a median of 8 minutes, much earlier than many out-of-hospital pharmacology studies.
  • Earlier out-of-hospital randomised trials of buffer therapy did not establish consistent long-term clinical benefit. ⁶⁷⁸⁹
  • Separate trials are still required if clinicians want definitive evidence for buffer therapy in prolonged out-of-hospital arrest or in highly selected acidotic subgroups.
• Observational evidence and resuscitation-time bias:
  • Observational studies of intra-arrest sodium bicarbonate are difficult to interpret because bicarbonate is often administered late as a rescue drug, creating resuscitation-time bias.
  • A Danish instrumental-variable analysis evaluated bicarbonate, calcium, and magnesium for in-hospital cardiac arrest before BIHCA and did not provide the same protection from confounding as randomisation. ¹⁰
  • Use trends and clinician surveys demonstrate persistence of the practice despite guideline discouragement, helping explain why a definitive randomised trial was needed. ¹¹¹²
• Individual treatment effects:
  • The editorial highlighted that clinicians treat individuals rather than populations and may continue to believe a subset benefits.
  • BIHCA’s prespecified subgroup analyses did not identify a clinically convincing subgroup.
  • More granular individual treatment-effect analyses may be hypothesis-generating, but any subgroup claim would require external validation or a dedicated trial.

Further Reading

• 2026Callaway CW. Will the Bicarbonate for In-Hospital Cardiac Arrest Trial Change Practice?
• 2026Lind PC, Hansen FG, Holmberg MJ, Jessen MK, Granfeldt A, Andersen LW. Bicarbonate and other buffer therapies in acute metabolic acidosis.
• 2025Drennan IR, Berg KM, Böttiger BW, et al. Advanced life support: 2025 ILCOR consensus on science with treatment recommendations.
• 2025Wigginton JG, Agarwal S, Bartos JA, et al. Part 9: Adult Advanced Life Support: 2025 American Heart Association Guidelines.
• 2025Soar J, Böttiger BW, Carli P, et al. European Resuscitation Council Guidelines 2025: adult advanced life support.
• 2024Ross CE, Sorcher JL, Gardner R, et al. Why physicians use sodium bicarbonate during cardiac arrest.
• 2023Holmberg MJ, Granfeldt A, Andersen LW. Bicarbonate, calcium, and magnesium for in-hospital cardiac arrest.
• 2019Moskowitz A, Ross CE, Andersen LW, Grossestreuer AV, Berg KM, Donnino MW. Trends over time in drug administration during adult in-hospital cardiac arrest.
• 2018Ahn S, Kim YJ, Sohn CH, et al. Sodium bicarbonate on severe metabolic acidosis during prolonged cardiopulmonary resuscitation.
• 2006Vukmir RB, Katz L; Sodium Bicarbonate Study Group. Sodium bicarbonate improves outcome in prolonged prehospital cardiac arrest.
• 1995Dybvik T, Strand T, Steen PA. Buffer therapy during out-of-hospital cardiopulmonary resuscitation.

Summary

• BIHCA randomised adults with in-hospital cardiac arrest who had received at least one dose of adrenaline to sodium bicarbonate up to 100 mmol versus placebo.
• Sustained ROSC was similar: 146/372 (39%) versus 150/407 (37%); adjusted RR 1.05; 95% CI 0.88 to 1.24; P=.62.
• Thirty-day survival and favourable neurological outcome were numerically higher with sodium bicarbonate but imprecise and not formally tested after the primary endpoint was not significant.
• Sodium bicarbonate increased pH and bicarbonate values after ROSC, confirming physiological effect, but did not improve early organ dysfunction or patient-centred outcomes.
• Alkalosis and hypernatraemia were more common with sodium bicarbonate, supporting avoidance of routine empiric administration.

Overall Takeaway

BIHCA is a practice-defining resuscitation pharmacology trial for adult in-hospital cardiac arrest: it shows that routine sodium bicarbonate after adrenaline changes blood gas values but not sustained ROSC. Its main practical message is to stop reflexive bicarbonate use during in-hospital cardiac arrest and reserve it for specific, biologically justified indications.

Bibliography

• 1.Callaway CW. Will the Bicarbonate for In-Hospital Cardiac Arrest Trial Change Practice? JAMA. Published online June 11, 2026.
• 2.Granfeldt A, Kirkegaard BL, Vallentin MF, et al; Bicarbonate for In-Hospital Cardiac Arrest (BIHCA) Investigators. Supplemental online content to: Sodium bicarbonate for in-hospital cardiac arrest: a randomized clinical trial. JAMA. Published online June 11, 2026.
• 3.Drennan IR, Berg KM, Böttiger BW, et al; Advanced Life Support Task Force Collaborators. Advanced life support: 2025 International Liaison Committee on Resuscitation consensus on science with treatment recommendations. Resuscitation. 2025;215(suppl 2):110806.
• 4.Wigginton JG, Agarwal S, Bartos JA, et al. Part 9: Adult Advanced Life Support: 2025 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation. 2025;152(suppl 2):S538-S577.
• 5.Soar J, Böttiger BW, Carli P, et al. European Resuscitation Council Guidelines 2025: adult advanced life support. Resuscitation. 2025;215(suppl 1):110769.
• 6.Lind PC, Hansen FG, Holmberg MJ, Jessen MK, Granfeldt A, Andersen LW. Bicarbonate and other buffer therapies in acute metabolic acidosis: a systematic review and meta-analysis. Acta Anaesthesiol Scand. 2026;70(4):e70212.
• 7.Dybvik T, Strand T, Steen PA. Buffer therapy during out-of-hospital cardiopulmonary resuscitation. Resuscitation. 1995;29(2):89-95.
• 8.Vukmir RB, Katz L; Sodium Bicarbonate Study Group. Sodium bicarbonate improves outcome in prolonged prehospital cardiac arrest. Am J Emerg Med. 2006;24(2):156-161.
• 9.Ahn S, Kim YJ, Sohn CH, Seo DW, Lim KS, Donnino MW, et al. Sodium bicarbonate on severe metabolic acidosis during prolonged cardiopulmonary resuscitation: a double-blind, randomized, placebo-controlled pilot study. J Thorac Dis. 2018;10(4):2295-2302.
• 10.Holmberg MJ, Granfeldt A, Andersen LW. Bicarbonate, calcium, and magnesium for in-hospital cardiac arrest: an instrumental variable analysis. Resuscitation. 2023;191:109958.
• 11.Moskowitz A, Ross CE, Andersen LW, Grossestreuer AV, Berg KM, Donnino MW. Trends over time in drug administration during adult in-hospital cardiac arrest. Crit Care Med. 2019;47(2):194-200.
• 12.Ross CE, Sorcher JL, Gardner R, et al. Why physicians use sodium bicarbonate during cardiac arrest: a cross-sectional survey study of adult and pediatric clinicians. Resusc Plus. 2024;20:100830.