Acute acidemia is common in critically ill adult patients. Use of sodium bicarbonate infusion for the treatment of severe metabolic acidemia is controversial and understudied.

Clinical Question

In critically ill patients with severe metabolic acidemia (pH ≤7⋅20), does the infusion of sodium bicarbonate, compared with no infusion, to reach and maintain a targeted pH of 7⋅30 decrease the primary composite out­come of mortality by day 28 or the presence of at least one organ failure at day 7?


389 adults (≥ 18 years of age) admitted within 48 hours to one of 26 ICUs in France from May 5,2015 through May 7, 2017 with pH ≤7⋅20, PaCO2 ≤45 mm Hg, and sodium bicarbonate concentration ≤20 mmol/L and with a total Sequential Organ Failure Assessment (SOFA) score of 4 or more or an arterial lactate concentration of 2 mmol/L or more.


125 to 250 mL of 4.2% sodium bicarb to target pH ≥ 7.3, MDD 1,000 mL


No sodium bicarbonate infusion


  • Primary: composite of death from any cause by 28 days after randomisation and the presence of at least one organ failure at 7 days after randomisation.
  • Secondary:
    • Use, duration, and number of days alive free of life­support interventions (renal­replacement therapy, mechanical ventilation, and vasopressors)
    • SOFA score at enrolment and at 1 day, 2 days, and 7 days after enrolment
    • Total fluid intake between enrolment and day 2
    • Adverse events of electrolytes that occurred during the ICU stay (plasma pH >7⋅45, hyperkalemia >5 mmol/L or <3⋅2 mmol/L; hypernatremia >145 mmol/L, and ionized calcemia <0⋅9 mmol/L)
    • Occurrence of ICU­acquired infection
    • Length of stay in the ICU


  • Prospective, multi-center, stratified randomization (accounting for age > 65, presence of acute kidney injury, presence of sepsis), controlled, unblinded trial. Authors argue infusing sodium bicarbonate raises pH, which will be apparent to treating physician, and as such blinding was not possible.
  • Intention-to-treat analysis was performed to compare intervention and control groups.
  • Indications for initiating renal replacement therapy were uniform across the study group
    • On admission, RRT was recommended when one or both of the following were present
      • Hyperkalemia > 6.5 mmol/L with EKG changes
      • Pulmonary edema with no urine output
    • Within 24 hours, RRT was recommended when two of the three criteria were present:
      • Urine output less than 0.3 mL/kg per hour for at least 24 hours
      • Arterial pH < 7.20 despite resuscitation
      • Hyperkalemia > 6.5 mmol/L


  • 109 patients had already received sodium bicarbonate at time of enrollment
  • 87 patients were in “terminal decline”
  • 76 patients had “treatment limitation”
  • 69 patients had chronic renal failure
  • 47 patients had immediate renal replacement therapy indication
  • 41 had ketoacidosis
  • 37 had digestive loss of sodium bicarbonate
  • 21 were eligible but not enrolled
  • 18 were included in another study
  • 13 had hyperkalemia with abnormal EKGs
  • 13 declined to participate
  • 11 were under guardianship protection
  • 11 withdrew consent

Primary Results

  • 389 patients randomized and maintained consent
    • 194 patients in control group
    • 195 patients in sodium bicarbonate group
  • Notably, 47 patients in the control group (47/194, 24.2%) received sodium bicarbonate. One patient assigned to the sodium bicarbonate group did not receive sodium bicarbonate. These patients were included in the intention-to-treat analysis.
  • Comparable demographic characteristics and illness severity between populations.
  • In intention-to-treat analysis:
    • No significant decrease in composite outcome in bicarbonate group (138/194 [71%] control group vs 128/195 [66%] bicarbonate group, p=0.24).
    • Significant reduction in need for RRT observed in bicarbonate group (100/194 [52%] control vs 68/195 [35%] bicarbonate, p = 0.0009, NNT =6).
    • In a subgroup analysis of patients with acute kidney injury score (AKIN) 2-3, a significant reduction in mortality was observed in the bicarbonate group (57/90 [63%] control vs 42/92 [46%] bicarbonate, p = 0.02, NNT = 6).


  • Large, multi-centered trial with clear inclusion criteria and study protocol.
  • A logical power analysis was performed to guide study enrollment.
  • Intention-to-treat and per-protocol analyses (supplementary data) had comparable findings.


  • Composite primary outcomes are, to a degree, inherently problematic. For instance, 28-day mortality and presence of organ dysfunction at 7-days are not equivalent outcomes to the patient. Here, the presence of a composite outcome may have weakened the signal.
  • Multiple secondary outcomes were measured and statistical analyses executed without performing Bonferroni correction for additional variables.
  • Several exclusion criteria exist.
  • 109 patients were excluded because they had already received sodium bicarbonate, another 24% of the control group received bicarbonate. This may have lessened the signal effect of bicarbonate.
  • The proportion of patients in whom the targeted pH of 7.30 was reached and maintained for at least 36 h from enrolment to day 2 was only 60%. The control group achieved this same target in 26% of patients
  • Physicians were not blinded and there was no control infusion. The authors provide explanation for this.
  • The protocol used 4.2% sodium bicarbonate (125–250 mL per infusion); a concentration that is not commonly used in the US. Use of “normal bicarb” 150 mEq in 1000 mL sterile water or D5W may not have the same effect. This is NOT crash cart bicarbonate (8.4%).
  • No data was reported for mechanical ventilation settings which may have influenced patients’ acid-base status.

Author's Conclusions

“In conclusion, in patients with severe metabolic acidemia, sodium bicarbonate treatment had no effect on the primary composite outcome (ie, mortality by day 28 or the presence of at least one organ failure at day 7) but decreased the need for renal replacement therapy. Additionally, sodium bicarbonate treatment did decrease mortality in the a­priori defined stratum of patients with acute kidney injury.”

Our Conclusions

In this study of critically ill adults with acidemia, sodium bicarbonate infusion had no effect on the primary outcome (mortality by day 28 or the presence of at least one organ failure at day 7) but did appear to have a large effect on reducing the need for RRT. In patients with AKIN 2-3, sodium bicarbonate infusion is associated with reduced mortality, however this was not a primary outcome or the main study population.

Potential Impact To Current Practice

Based on this study, it would be reasonable to consider sodium bicarbonate infusion in a select group of critically ill patients with metabolic acidosis (pH ≤7⋅20, PaCO2 ≤45 mm Hg, and sodium bicarbonate concentration ≤20 mmol/L), especially those with acute kidney injury class 2-3.

Bottom Line

Bicarbonate does not replace the need to correct the underlying cause of the metabolic acidosis. pH guided fluid resuscitation, particularly sodium bicarbonate infusion, may reduce need for RRT in critically ill patients with metabolic acidosis, and may have an effect on mortality in a subset of acidotic patients with AKIN 2-3. Further research is warranted, particularly with sodium bicarbonate concentrations more commonly used in the United States, to determine applicability and efficacy of the intervention in our practice scenario.

Read More