Bicarbonate May Reduce Dialysis Needs in Critically Ill Patients

A comprehensive Bayesian systematic review and meta-analysis of randomised controlled trials has evaluated the clinical effects of sodium bicarbonate therapy for critically ill adults suffering from metabolic acidosis. Published in Intensive Care Medicine, the study synthesised data from four randomised controlled trials involving a total of 1,544 patients. The primary objective was to determine whether administering intravenous sodium bicarbonate improves mortality rates or renal outcomes compared with placebo or usual care. The analysis revealed that while a clear survival benefit was not demonstrated, there was a high posterior probability that the therapy reduces the need for renal replacement therapy.

Metabolic acidosis is a frequent complication in intensive care units, often associated with severe organ dysfunction and increased mortality. In its most severe forms, acidaemia can impair myocardial contractility, promote vasodilation, and exacerbate haemodynamic instability, particularly in patients already experiencing shock or acute kidney injury. Although treating the underlying cause of the acidosis remains the paramount clinical priority, the utility of actively correcting the acid-base balance has remained a subject of significant uncertainty. Sodium bicarbonate administration increases extracellular pH, but it also carries physiological risks, including increased carbon dioxide generation, sodium load, and fluid burden.

These competing physiological effects have created persistent heterogeneity in clinical practice. Contemporary guidance, such as the Surviving Sepsis Campaign, generally suggests against routine bicarbonate administration for haemodynamic support in hypoperfusion-induced lactic acidaemia. Instead, its use is reserved mainly for patients with severe metabolic acidaemia and concomitant acute kidney injury. Previous randomised trials, including BICAR-ICU and BICARICU-2, provided initial insights but did not fully resolve the uncertainty regarding the balance of benefits and harms or clearly identify the specific patient populations most likely to benefit from the intervention.

The current study employed a Bayesian random-effects meta-analysis to evaluate the effects of sodium bicarbonate on mortality and renal outcomes. This statistical framework allows for the direct estimation of the probability of benefit across clinically relevant outcomes, providing a probabilistic interpretation that may facilitate more nuanced clinical decision-making. The researchers searched MEDLINE, Embase, and CENTRAL databases from inception to December 2025, with an updated search conducted on 2 April 2026. The inclusion criteria were strict, limited to randomised controlled trials comparing intravenous sodium bicarbonate with placebo or usual care in adult intensive care unit patients with metabolic acidosis.

Four randomised controlled trials met the inclusion criteria, enrolling 1,544 patients with 768 assigned to sodium bicarbonate and 776 to the control group. These trials were conducted across intensive care units in France, Australia, New Zealand, and Japan. The study designs varied, with BICAR-ICU and BICARICU-2 being open-label, while the SODa-BIC pilot and SODa-BIC trial were double-blind and placebo-controlled. Eligibility criteria also differed; BICAR-ICU and BICARICU-2 enrolled patients with a pH of 7.20 or lower, whereas the SODa-BIC trials enrolled patients with a pH below 7.30 and a base excess of -4 mEq/L or lower. BICARICU-2 specifically enrolled patients with stage 2 to 3 acute kidney injury, while the other trials included broader intensive care populations.

The primary outcome of the analysis was short-term all-cause mortality, defined as mortality at 28 or 30 days after randomisation. The results indicated that sodium bicarbonate did not show a clear short-term mortality benefit. The pooled risk ratio was 0.94, with a 95% credible interval ranging from 0.73 to 1.22. The posterior probability of a mortality benefit was 72.1%, which, while suggesting a likelihood of benefit, remained compatible with both benefit and harm due to the wide credible interval. Findings were broadly similar across different prior specifications, indicating the robustness of the mortality estimates within the Bayesian framework.

In contrast to the mortality findings, the analysis revealed a strong association between sodium bicarbonate therapy and reduced initiation of renal replacement therapy. The risk ratio for renal replacement therapy initiation was 0.74, with a 95% credible interval of 0.59 to 0.98. The posterior probability of reducing renal replacement therapy use was 97.2%, indicating a high degree of certainty that the therapy lowers the risk of requiring dialysis. These findings were consistent across different prior specifications and were supported by a post hoc analysis showing that the probability of between-group differences in time to renal replacement therapy initiation being within 12 hours was 85.4%.

Secondary outcomes, including 90-day mortality, vasopressor-free days, renal replacement therapy-free days, and intensive care unit and hospital length of stay, showed estimates centred near no effect with substantial uncertainty. The credible intervals for these outcomes included no effect, suggesting that sodium bicarbonate does not significantly impact these broader clinical measures. However, a post hoc analysis indicated an 82.7% probability of reducing renal replacement therapy dependence at the longest available follow-up, although the credible interval for this specific outcome also included no effect. The certainty of evidence was rated as moderate for short-term mortality, renal replacement therapy initiation, and 90-day mortality, and high for intensive care unit and hospital length of stay.

The biological plausibility of a reduction in renal replacement therapy without a corresponding mortality benefit is supported by the physiological effects of bicarbonate. Sodium bicarbonate directly corrects extracellular acidaemia, which may reduce the need for renal replacement therapy when severe acid-base disturbance is a primary indication for dialysis. Correction of acidaemia may also improve haemodynamic stability and hyperkalaemia, allowing time for the recovery of kidney function in some patients. This could enable spontaneous renal recovery in a subset of patients who would otherwise have received dialysis, thereby avoiding the procedure altogether rather than simply delaying its initiation.

Subgroup analyses suggested that the probability of benefit was greater among patients with more advanced acute kidney injury and more severe acidaemia. These findings align with previous analyses and support a selective, phenotype-informed treatment strategy rather than routine use across all patients with metabolic acidosis. However, the estimates were imprecise and should be interpreted with caution. The study acknowledged several limitations, including the small number of available trials, which limited precision for several outcomes and heterogeneity estimates. The included trials enrolled clinically distinct patient populations, which may limit the generalisability of the pooled estimates.

Additionally, two of the four trials were open-label, raising the possibility that the observed reduction in renal replacement therapy, a clinician decision-dependent outcome, may have been influenced by knowledge of treatment allocation. Even in double-blind trials, complete masking may be challenging for an intervention guided by acid-base physiology, as changes in pH, bicarbonate, and other markers may allow clinicians to infer treatment allocation. The meta-analysis was based on aggregate trial-level data rather than individual participant data, preventing a detailed evaluation of treatment effects according to specific patient-level characteristics. The small number of trials also limited the assessment of publication bias.

From a clinical perspective, the current evidence does not support the routine use of sodium bicarbonate for all critically ill patients with metabolic acidosis when the primary goal is to improve survival. However, the consistently high probability of reducing renal replacement therapy use suggests a clinically meaningful effect on kidney-related outcomes. Given the burden, risks, and costs associated with dialysis in the intensive care unit, this reduction may be significant. The findings support a selective approach, with bicarbonate most plausibly considered in patients with severe acidaemia and concomitant acute kidney injury, particularly when acid-base derangement contributes to haemodynamic instability or may precipitate dialysis.

Future trials should focus on higher-risk populations with severe acidaemia and advanced acute kidney injury, apply standardised criteria for renal replacement therapy initiation, and evaluate kidney-specific and longer-term outcomes. Ongoing trials, such as MOSAICC, are expected to provide further evidence on the clinical and cost-effectiveness of sodium bicarbonate in critically ill patients. Individual-patient data meta-analyses may further clarify whether treatment effects differ according to baseline pH, acute kidney injury severity, or sepsis status, helping to refine patient selection for therapy. Careful monitoring for adverse metabolic effects, including hypernatraemia, metabolic alkalosis, and hypocalcaemia, remains essential for any patient receiving this treatment.

Subject of Research: Bayesian meta-analysis of sodium bicarbonate therapy for metabolic acidosis in critically ill patients

Article Title: Sodium bicarbonate therapy for metabolic acidosis in critically ill patients: a Bayesian systematic review and meta-analysis of randomised controlled trials

Article References: Kitisin, N., Walker, H. G. M., Jung, B., Jaber, S., Mouncey, P., Ke, L., Raykateeraroj, N., Hikasa, Y., Caroli, A., & Serpa Neto, A. (2026). Sodium bicarbonate therapy for metabolic acidosis in critically ill patients: a Bayesian systematic review and meta-analysis of randomised controlled trials. Intensive Care Medicine. https://doi.org/10.1007/s00134-026-08602-3

Image Credits: AI Generated

DOI: 10.1007/s00134-026-08602-3

Keywords: sodium bicarbonate, metabolic acidosis, critical care, Bayesian meta-analysis, renal replacement therapy, acute kidney injury, intensive care medicine, randomised controlled trials, Sodium, bicarbonate, therapy, metabolic