Emergency Medicine · Medical Literature
Sodium Bicarbonate for Critically Ill Adults with Metabolic Acidosis and Shock
Abstract
New England Journal of Medicine, Ahead of Print.
Clinical relevance
In critically ill adults with vasopressor-dependent shock and metabolic acidosis (pH <7.30), does a short-course intravenous sodium bicarbonate infusion titrated to normalize pH and base excess reduce major adverse kidney events within 30 days compared with placebo? __SNAPSHOT__ {"studyDesign":"Pragmatic, adaptive, double-blind, randomized controlled trial (superiority design; SODa-BIC trial)","setting":"55 ICUs across seven countries (Australasia, Brazil, Japan, Canada, and Ireland); international multicenter","sampleSize":"n=500 (sodium bicarbonate: n=245; placebo: n=255)","population":"Adults ≥18 years in the ICU with metabolic acidosis (pH <7.30, base excess ≤−4 mmol/L, PaCO₂ ≤45 mmHg if non-intubated or ≤50 mmHg if intubated) receiving a continuous vasopressor infusion. Approximately 40% had septic shock; ~25% had metabolic acidosis attributed to cardiogenic shock or cardiac arrest; ~80% were mechanically ventilated; most had mild AKI at baseline (creatinine ~1.5 mg/dL).","intervention":"Intravenous sodium bicarbonate (5% solution, 600 mmol/L) infused for up to 5 hours, with rate titrated to achieve a target pH ≥7.30 and base excess ≥0 mmol/L","comparator":"Placebo (5% dextrose in water) infused identically for up to 5 hours; open-label sodium bicarbonate bolus was permitted in both arms if clinically indicated by the treating team","primaryOutcome":"Major adverse kidney event within 30 days (MAKE30): a composite of all-cause death, new receipt of renal replacement therapy, or persistent renal dysfunction (serum creatinine ≥200% of baseline), censored at hospital discharge or day 30, whichever occurred first","followUp":"30 days (primary outcome); secondary outcomes assessed at 30 and 90 days","fundingSource":"Government (National Health and Medical Research Council of Australia; ClinicalTrials.gov NCT05697770); no industry funding reported"}
Key findings
The primary outcome of MAKE30 occurred in 98/244 patients (40.2%) in the sodium bicarbonate group versus 100/254 patients (39.4%) in the placebo group (adjusted absolute difference +1.2 percentage points; 95% CI −7.1 to +9.4; P=0.78), demonstrating no statistically significant benefit. In-hospital mortality by day 30 was 25.4% with sodium bicarbonate versus 24.0% with placebo (adjusted difference +1.8 percentage points; 95% CI −5.6 to +9.2). Renal replacement therapy was used in 16.8% of the bicarbonate group versus 20.9% of the placebo group (adjusted difference −3.9 percentage points; 95% CI −10.6 to +2.7), a numerically lower rate that did not reach statistical significance. Persistent renal dysfunction occurred numerically less often in the bicarbonate group (14% vs. 18%), and dialysis dependence at day 30 was also numerically lower (5.7% vs. 9.1%); neither difference was statistically significant. Sodium bicarbonate did achieve biochemical separation: during the first 24 hours, the bicarbonate group had pH values approximately 0.05 units higher with higher serum bicarbonate than placebo — confirming physiological effect without clinical benefit on hard outcomes. Regarding safety, four patients (1.6%) in the sodium bicarbonate group experienced an adverse effect (significant hypokalemia <3.0 mmol/L requiring replacement) versus none in the placebo group (P=0.06). NNT for MAKE30 prevention: not calculable (primary outcome was neutral). NNH for hypokalemia-type adverse event: approximately 63 (calculated from 4/244 vs. 0/254; absolute risk increase ~1.6%).
Practice implications
⚠ Only the abstract and select paywalled data points reported in secondary sources are available; the full dataset, subgroup analyses, and pre-specified sensitivity analyses have not been independently reviewed. This limitation should be considered when applying these findings. This trial established that routine short-course intravenous sodium bicarbonate, titrated to normalize extracellular pH over ≤5 hours, does not reduce MAKE30, mortality, or need for RRT in vasopressor-dependent ICU patients with moderate metabolic acidosis (pH <7.30). The trial confirmed robust biochemical efficacy — bicarbonate corrected pH and base excess as intended — yet this did not translate into patient-centered benefit, consistent with the 'treating the number' concern articulated around Goodhart's Law: pH normalization as a surrogate target did not correspond to meaningful clinical improvement. The findings apply to the tested phenotype: vasopressor-dependent adults with moderate metabolic acidosis and predominately mild-to-moderate AKI. The trial does NOT establish or refute benefit of bicarbonate in: (1) severe metabolic acidosis with concurrent moderate-to-severe AKI (the subgroup that showed signal in BICAR-ICU); (2) bicarbonate-wasting etiologies (renal tubular acidosis, diarrhea-related hyperchloremic acidosis); (3) toxicological indications (e.g., salicylate poisoning, tricyclic overdose); (4) severe hyperkalemia; or (5) extreme pH emergencies. These scenarios were either excluded or underrepresented. Observational data had suggested a modest mortality reduction (absolute risk reduction ~1.9%, RR 0.86), but this trial's null result reinforces that residual confounding likely inflated those estimates. The permitted open-label bicarbonate crossover in both arms may have attenuated between-group separation beyond 5 hours, a potential dilution bias. Secondary and subgroup analyses were not multiplicity-adjusted and should be considered hypothesis-generating only. __LIMITATIONS__ Trial powered to detect a 14 percentage-point absolute reduction in MAKE30 (a large effect size); smaller but clinically meaningful differences (e.g., 5–7 percentage points) would not have been detectable with 90% power at this sample size.,Open-label bicarbonate was permitted in both groups if clinically indicated, potentially diluting between-arm differences beyond the 5-hour infusion window.,The MAKE30 composite endpoint weights death, RRT use, and persistent renal dysfunction equally; RRT initiation is a clinician-dependent decision and subject to practice variability even with double-blinding of bedside staff.,Generalizability is uncertain for patients with severe AKI at baseline, non-lactic forms of metabolic acidosis, or extreme pH (<7.10), who may represent a distinct biological phenotype not well-represented in this trial.,Secondary and subgroup analyses were not adjusted for multiple comparisons and must be treated as exploratory.,30-day follow-up captures short-term kidney outcomes; longer-term renal recovery, dialysis dependence beyond 90 days, and neurocognitive outcomes were not reported in available abstract data.,Although funded by a government body with no reported industry involvement (a strength), the pragmatic design and international heterogeneity in vasopressor thresholds, RRT initiation practices, and concomitant fluid strategies may introduce uncontrolled variation. __BOTTOM_LINE_CATEGORY__ potentially_practice_informing __BOTTOM_LINE__ In vasopressor-dependent critically ill adults with metabolic acidosis (pH <7.30), a short-course sodium bicarbonate infusion titrated to normalize pH improved blood gas parameters but did not reduce the 30-day composite of death, renal replacement therapy, or persistent renal dysfunction compared with placebo (MAKE30: 40.2% vs. 39.4%; P=0.78). This well-powered, double-blind trial provides high-quality evidence against routine bicarbonate administration for pH correction in this population, though it does not address established or rescue indications such as severe AKI-associated acidosis, bicarbonate-wasting states, or toxicological emergencies. Clinicians should refocus on treating the underlying cause of acidosis rather than targeting extracellular pH normalization as a primary goal.