Sodium bicarbonate did not improve survival after in-hospital cardiac arrest

Sodium bicarbonate, a buffer that clinicians have long pushed into the veins of patients whose hearts have stopped, did not improve the chance that circulation returned — and did not improve survival — in a randomised, double-blind, placebo-controlled trial of adults with in-hospital cardiac arrest [s1]. The drug is given on the reasoning that a stopped heart floods the body with acid, and that correcting the acid should help resuscitation work. The trial found no such benefit, and instead documented the metabolic disturbances the drug is known to cause [s1].

What the trial tested

The trial was run at 21 hospitals in Denmark and enrolled adults who had an in-hospital cardiac arrest and had already received at least one dose of adrenaline during resuscitation [s1]. Patients were randomly assigned to receive up to 100 mmol of intravenous sodium bicarbonate or a matching placebo, with neither the treating team nor the investigators knowing which [s1]. Enrolment ran from 6 February 2023 to 11 February 2026, with the final 90-day follow-up completed on 4 May 2026 [s1]. The study was investigator-initiated and led from the University of Aarhus, not by a drug company — a point that matters because bicarbonate is generic, off-patent and has no commercial sponsor with a reason to test it [s2].

The primary outcome was sustained return of spontaneous circulation — a restarted heartbeat that holds. The two most important secondary outcomes were being alive at 30 days, and being alive at 30 days with a favourable neurologic outcome, defined as a score of 0 to 3 on the modified Rankin Scale, a 0-to-6 measure on which higher scores mean greater disability [s1].

What it found

Of 2,913 patients screened, 913 were randomised, and 779 were eligible for the primary analysis — 372 assigned to bicarbonate and 407 to placebo [s1]. Their median age was 73 years (interquartile range 64 to 79) and 502 (64%) were male [s1].

Sustained return of circulation occurred in 146 patients (39%) given bicarbonate and 150 (37%) given placebo — a risk ratio of 1.05 (95% confidence interval 0.88 to 1.24; P = .62), which is statistical language for no detectable effect [s1]. Survival pointed the same way. At 30 days, 45 patients (12%) in the bicarbonate group and 37 (9.1%) in the placebo group were alive (risk ratio 1.25; 95% CI 0.84 to 1.88), and a favourable neurologic outcome was reached by 30 patients (8.1%) versus 22 (5.4%) (risk ratio 1.39; 95% CI 0.82 to 2.34) [s1]. Those survival intervals are wide and cross 1, meaning the trial cannot rule a modest benefit in or out — but neither did it show one, and the primary endpoint was flatly negative [s1].

What the drug did do was disturb the blood: alkalosis and hypernatremia — an overshoot into alkalinity and a rise in sodium — were more common in the bicarbonate group after the arrest [s1]. These are the predictable costs of infusing a strong buffer, and here they came with no offsetting benefit [s1].

How to read it

The result is a textbook negative trial, and its authors state the practical implication plainly: the findings do not support routine administration of sodium bicarbonate for patients with in-hospital cardiac arrest [s1]. That conclusion carries weight precisely because of the trial's design — double-blind, placebo-controlled and adequately sized to answer its primary question, which is uncommon for a resuscitation intervention that entered practice on physiological reasoning rather than evidence.

Two limits are worth stating. The trial studied in-hospital arrests in one country's health system, where the causes of arrest and the speed of response differ from the out-of-hospital setting, so the result should not be stretched to every situation [s1]. And it tested routine, unselected use; it did not address the specific circumstances — such as a known severe pre-existing acidosis, a dangerously high potassium level, or certain drug overdoses — in which bicarbonate is given for a defined reason rather than reflexively [s1]. The trial's message is about the reflex, not those targeted uses.

What to watch

The open question is whether guideline committees, many of which already restrict bicarbonate to specific indications, tighten that language further in light of a clean negative trial [s1]. The broader pattern is familiar: resuscitation medicine is steadily subjecting its inherited habits to placebo-controlled tests, and several once-standard reflexes have not survived the encounter. This article describes research and is not medical advice; decisions during resuscitation are made by treating clinicians.

Sources

• Sodium Bicarbonate for In-Hospital Cardiac Arrest: A Randomized Clinical Trial — JAMA, 2026
• Bicarbonate for In-Hospital Cardiac Arrest — A Randomized, Double-Blind, Placebo-Controlled Trial (NCT05564130) — ClinicalTrials.gov