Date: November 12, 2026

Guest Skeptic: Dr. Gillian Schmitz is a board-certified Emergency Physician practicing at the Naval Medical Center in San Diego. She is also a former President of the American College of Emergency Physicians (ACEP). Gillian was the guest skeptic for the SGEM Xtra: Now and Forever – A League of Their Own (Women in EM).

Reference: Granfeldt et al. Bicarbonate for In-Hospital Cardiac Arrest (BIHCA) Investigators. Sodium Bicarbonate for In-Hospital Cardiac Arrest: A Randomized Clinical Trial. JAMA 2026 Aug

Case: A 73-year-old man presents to the emergency department (ED) with increasing shortness of breath and generalized weakness. His past medical history includes hypertension, diabetes, coronary artery disease, atrial fibrillation, and chronic heart failure. Shortly after being placed on a monitored stretcher, he suddenly becomes unresponsive. He has no palpable pulse. CPR is started immediately. The monitor demonstrates pulseless electrical activity (PEA). There is no documented do-not-resuscitate order. He receives epinephrine approximately five minutes after the arrest begins. Despite continued high-quality CPR, ventilation, vascular access, and evaluation for reversible causes, he remains pulseless after the first dose of epinephrine. Someone on the resuscitation team asks: “Should we give an amp of bicarb?”

Background: In-hospital cardiac arrest (IHCA) is not rare. The 2025 American Heart Association guidelines estimate that roughly 1 in 100 hospitalized adults experiences a cardiac arrest, representing about 292,000 adult IHCAs each year in the United States. Around 72% achieve return of spontaneous circulation (ROSC), but only about 24% survive to hospital discharge. Among those who survive, neurological outcomes are considerably better than overall survival rates might suggest.

IHCA is not simply an out-of-hospital cardiac arrest that happens in the hospital. Patients are usually sicker, comorbidities are common, and deterioration may already be recognized. These patients are often witnessed arrests because they are on cardiac monitors. Trained personnel, defibrillators, intravenous access, medications, and advanced airway equipment can potentially be at the bedside within minutes. Approximately 60% of adult IHCAs occur in higher-acuity locations such as the intensive care unit (ICU), ED, or operating room (OR), with the remainder occurring on general inpatient units.

Treatment starts with the fundamentals: recognize the arrest, provide high-quality CPR, identify the rhythm, and promptly defibrillate ventricular fibrillation or pulseless ventricular tachycardia. Epinephrine remains in the algorithms, despite a lack of evidence, while clinicians simultaneously search for reversible causes. Once ROSC hopefully occurs, attention turns to oxygenation, ventilation, hemodynamics, identifying the cause of arrest, coronary reperfusion when appropriate, and post-resuscitation care.

Then there are the medications that seem physiologically reasonable. Sodium bicarbonate is a good example. Cardiac arrest produces profound metabolic and respiratory derangements, so correcting acidosis sounds attractive. The problem is that improving a laboratory-oriented outcome (LOO) is not necessarily the same thing as improving a patient-oriented outcome (POO). Current AHA guidance recommends against routine sodium bicarbonate during cardiac arrest, while preserving a role in selected circumstances such as certain sodium-channel blocker poisonings and potentially specific electrolyte emergencies. ,


Clinical Question: In adults with in-hospital cardiac arrest who have received epinephrine, does intravenous sodium bicarbonate, compared with placebo, increase sustained return of spontaneous circulation and improve survival or neurological outcomes?


Reference: Granfeldt et al. Bicarbonate for In-Hospital Cardiac Arrest (BIHCA) Investigators. Sodium Bicarbonate for In-Hospital Cardiac Arrest: A Randomized Clinical Trial. JAMA 2026 Aug

  • Population: Adults aged 18 years or older experiencing an IHCA who received at least one dose of epinephrine during the resuscitation.
    • Exclusions: Clearly documented Do Not Resuscitate (DNR) order before arrest; previous enrollment in the trial; extracorporeal circulation or LVAD support at the time of arrest; known or suspected pregnancy; known patient objection to trial participation; or a clinical indication for sodium bicarbonate.
  • Intervention: 50 mL of 8.4% sodium bicarbonate, equivalent to 50 mmol, given as soon as possible after the first dose of epinephrine. If the patient remained in arrest after the subsequent epinephrine dose, another 50 mmol could be given, for a maximum of 100 mmol.
  • Comparison: Matching placebo consisting of 50 mL of 0.9% sodium chloride supplied in identical vials.
  • Outcomes:
    • Primary Outcome: Sustained ROSC, defined as signs of circulation without further chest compressions for at least 20 minutes.
    • Secondary Outcomes: Key secondary outcomes were 30-day survival and 30-day survival with a favourable neurological outcome, defined as modified Rankin Scale 0–3. Other outcomes included 90-day survival, neurological outcome, EQ-5D-5L quality of life, SOFA scores, laboratory measurements, hospital disposition, cause of death, and prespecified adverse events including alkalosis, hypernatremia, severe hypocalcemia, hypokalemia, and markedly elevated lactate.
  • Type of Study: This was a multicentre, randomized, parallel-group, double-blind, placebo-controlled superiority randomized clinical trial.

Authors’ Conclusions: “There was no significant difference in sustained return of spontaneous circulation between sodium bicarbonate and placebo in adults with in-hospital cardiac arrest. These findings do not support routine administration of sodium bicarbonate for patients with in-hospital cardiac arrest.”

Quality Checklist for Randomized Clinical Trials:

  1. Did the study population include or focus on those in the emergency department? Yes
  2. Were the patients adequately randomized? Yes
  3. Was the randomization process concealed? Unsure
  4. Were patients analyzed in the groups to which they were randomized? No
  5. Were patients recruited consecutively? No
  6. Were both groups similar with respect to prognostic factors? Yes
  7. Were patients, clinicians and outcome assessors blinded? Yes
  8. Were groups treated equally except for the intervention? Yes
  9. Was follow-up complete? Yes
  10. Were all patient-important outcomes considered? Yes
  11. Was the treatment effect large and precise enough to be clinically significant? No
  12. Who funded the trial? Novo Nordisk Foundation. The publication states that the foundation had no role in trial design, conduct, analysis, interpretation, manuscript preparation, or publication decision.
  13. Did the authors declare conflicts of interest? Several investigators disclosed grants or lecture fees from pharmaceutical organizations outside the submitted work.

Results: The primary analysis included 779 patients: 372 assigned to sodium bicarbonate and 407 to placebo. Median age was 73 years, and 64% were male. Most arrests occurred on a standard hospital ward (64%), while about 16% occurred in the ED. A large majority (88%) had an initial nonshockable rhythm, predominantly PEA or asystole. Approximately three-quarters of arrests were witnessed, and roughly 42% were monitored. Median time from arrest to epinephrine was 5 minutes, and median time to trial drug was 8 minutes. Known metabolic acidosis before arrest was documented in only about 14%.


Key Results: Giving sodium bicarbonate during in-hospital cardiac arrest did not significantly increase sustained ROSC compared with placebo.


  • Primary Outcome: Sustained ROSC occurred in 39% patients assigned bicarbonate versus 37% assigned placebo: RR 1.05 (95% CI 0.88–1.24), P=0.62.
  • Secondary Outcomes: No statistical difference in 30-day or 90-day survival or survival with favourable neurological function
    • 30-day survival was 12% vs 9.1% (RR 1.25, 95% CI 0.84–1.88)
    • 30-day survival with favourable neurological function occurred in 8.1% vs 5.4% (RR 1.39, 95% CI 0.82–2.34).
    • 90-days survival was 11% versus 8.1%
    • 90-day favourable neurological survival was 8.6% versus 6.9%
    • Quality of life and organ dysfunction showed no clear group differences
    • Post-arrest pH and bicarbonate were higher after bicarbonate, while alkalosis occurred in 35% versus 20% and hypernatremia in 42% vs 29%

1. Modified Intention-To-Treat (mITT): The most important internal-validity concern is the removal of randomized participants after allocation. Of 913 randomized patients, only 779 were included in the primary analysis. The authors argue that because treatment allocation remained blinded, these exclusions should be unbiased. That is reassuring, but high-quality methodology favours keeping randomized participants in their assigned groups because post-randomization exclusions can undermine randomization’s protection against bias. An ITT analysis is particularly valuable because it preserves the prognostic balance generated by randomization.

2. Potential Selection Bias: A very large number of patients were not enrolled for logistical reasons. These included clinical teams forgetting the study, patients achieving ROSC before administration, resuscitation being terminated, staff being unavailable, or study medication being unavailable. These omissions are understandable in the chaos of cardiac arrest research, but they mean this was not a straightforward consecutive sample of eligible IHCAs. If enrolled patients differ systematically from missed patients in prognosis or resuscitation circumstances, representativeness suffers, and the estimated treatment effect may be biased.

3. Powered for Return of Spontaneous Circulation (ROSC): Sustained ROSC is a clinically meaningful outcome. You can’t achieve survival with a favourable neurologic outcome without first restoring a pulse. However, it remains more of an intermediate disease-oriented outcome (DOO). Patients really care about being alive and able to return to an acceptable quality of life (POO). The confidence interval around the ROSC result makes a large effect unlikely, but the trial was not sufficiently powered to exclude clinically meaningful differences in 30-day survival or favourable neurological survival.

4. Biological Heterogeneity: The theoretical reason for administering bicarbonate is largely related to acid-base physiology, yet intra-arrest pH and bicarbonate could not be measured before enrolment. Only a minority had documented pre-arrest metabolic acidosis. Consequently, the trial combines patients with potentially very different degrees and causes of acidemia. A treatment that is ineffective across a heterogeneous arrest population could conceivably have a different effect in a biologically defined subgroup. The authors appropriately acknowledge this limitation.

5. External Validity: This was a relatively large Danish hospital trial involving patients whose arrests persisted long enough to require epinephrine. Nearly nine in ten had nonshockable rhythms, only about one in six arrested in the ED, and people with an established clinical indication for bicarbonate were explicitly excluded. The findings therefore apply best to routine empiric bicarbonate during epinephrine-requiring IHCA. They don’t apply to every cardiac arrest, rapidly terminated shockable arrests, sodium-channel blocker poisoning, severe hyperkalemia, or every ED resuscitation system around the world.

Comment on the Authors’ Conclusion Compared to the SGEM Conclusion: We generally agree with the authors’ conclusions.


SGEM Bottom Line: For routine epinephrine-requiring adult IHCA, bicarbonate seems better at changing chemistry than changing outcomes. Stick with excellent CPR, defibrillation when indicated, and aggressively treat reversible causes.


Case Resolution: The 73-year-old patient remains in PEA after epinephrine. There is no evidence of hyperkalemia, sodium-channel blocker toxicity, or another specific indication for bicarbonate. Sodium bicarbonate is not given simply because the resuscitation has been going for several minutes or because the patient is presumed to be acidotic. Continue high-quality CPR with minimal interruptions, optimize ventilation without hyperventilating, and we aggressively work through reversible causes.

Dr. Gillian Schmitz

Clinical Application: For clinicians already following contemporary ACLS recommendations and not routinely administering sodium bicarbonate, there is little reason to change practice. This trial provides considerably stronger direct evidence supporting that approach. For clinicians who routinely give an amp or two of bicarbonate to a prolonged arrest simply because “they must be acidotic by now,” this should make that practice harder to defend. The drug produced the expected biochemical effects and more alkalosis and hypernatremia without demonstrating improved ROSC. The conclusion should not be never to give bicarbonate in cardiac arrest. The study excluded patients who already had an indication for it. A better conclusion is: do not give it routinely without a reason.

What Do I tell the Patient? Since the patient during cardiac arrest cannot participate in the discussion, this conversation will usually happen with the survivor or family afterward.

Keener Kontest: Last week’s winner was Dr. Steven Stelts. He knew the American College of Emergency Medicine was founded on 16 August 1968 in Lansing, Michigan, by a group of eight emergency physicians. The first president was John G. Wiegenstein.

Other SGEM Episodes on IHCAs:

  • SGEM#50: Under Pressure – Vasopressin, Steroids and Epinephrine in Cardiac Arrest
  • SGEM#197: Die Trying – Intubation of In-Hospital Cardiac Arrests
  • SGEM#350: How Did I Get Epi Alone? Vasopressin and Methylprednisolone for In-Hospital Cardiac Arrests
  • SGEM#394: Say Bye Bye Bicarb for Pediatric In-Hospital Cardiac Arrest

Remember to be skeptical of anything you learn, even if you heard it on the Skeptics’ Guide to Emergency Medicine.