Sodium bicarbonate infusions: do they improve outcomes, or just the pH? (Review)

Putting SODa-Bic, BICARICU-1 and 2 into the acid bath

In the 1990s, an elite rower had a recorded venous pH of 6.74 after a 2,000-meter race. Rather than collapsing or requiring hospitalization, he caught his breath, chatted with the crowd, and drove home.

Elite sprinters routinely hit pHs of 7.10 with lactate levels exceeding 10 mmol/L. Even weekend-warrior marathoners can easily enter the pH range of 7.20 to 7.30 near the end of a race.

So who’s afraid of a few extra stray hydrogen ions?

Well, the etiology matters: exertional metabolic acidosis is brief, self-limited, and occurring in healthy people with intact cardiopulmonary-renal function. By contrast, critically ill patients have ongoing acid generation from organ dysfunction with diminishing physiologic reserves. Acidemia can depress myocardial function and blunt catecholamine responsiveness, although the thresholds at which this occurs have been debated.

But many critically ill patients with acidemia have something distantly in common with those elite athletes: their acidemia is an adaptive response to stress, not necessarily harmful, and will resolve without consequence after the underlying stressors are relieved.

Others have severe acidemia from an unrelenting metabolic decompensation near death. For them, buffering out some H+ may make their labs look marginally better for a short time, but will have zero impact on the ultimate outcome.

Is there a third group, the Goldilockses of the ICU, who have acidemia that is both maladaptive (i.e., worsening an already bad situation) and remediable?

A series of randomized trials have sought to find this elusive group: BICARICU-1 and BICARICU-2, which we reviewed here —

Sodium Bicarbonate Administration in Severe Metabolic Acidemia: the BICAR-ICU Trial

Bicarb for AKI + acidosis prevents dialysis needs in the ICU (BICARICU-2 trial)

— and most recently the SODa-BIC trial in NEJM.

The SODa-BIC Trial

In 55 ICUs in seven countries (in Australasia, Brazil, Japan, Canada, and Ireland), 500 critically ill adults with vasopressor-dependent shock and metabolic acidosis (pH < 7.30) were enrolled.

It was a sick group. About 40% had septic shock; ~25% had metabolic acidosis attributed to cardiogenic shock or cardiac arrest. Most had mild acute kidney injury (Cr ~1.5 mg/dL). About 80% were mechanically ventilated.

They were randomized to receive either 5% sodium bicarbonate (600 mmol/L) or placebo (5% dextrose) infusions for ≤5 hours to achieve pH ≥7.30, and followed for 30 days to assess the incidence of a “major adverse kidney event”: either death, need for dialysis or CRRT, or a persistent doubling (or greater) of baseline creatinine. (This composite endpoint, reused across trials, is referred to as MAKE30.)

SODa-BIC Results:

There was separation between groups, and bicarbonate improved acidosis. During the first 24 hours, the sodium bicarbonate group had pH values ~0.05 higher with serum bicarbonate levels than placebo.

But after 30 days, the rate of the MAKE30 composite outcome was no different between arms — 40.2% in the NaHCO3 group and 39.4% in the placebo group.

In-hospital mortality was near-identical (25.4% for NaHCO3 vs 24.0% with placebo).

However, sodium bicarbonate-treated patients had persistent renal dysfunction less often, numerically (14% vs 18% with placebo), required dialysis/CRRT 4% less often overall (16.8% vs 20.9%) and were less often dialysis-dependent at 30 days (5.7% vs 9.1%)—none of which were statistically significant.

There were no significant safety events, but four patients in the sodium bicarb group had significant hypokalemia (<3.0 mmol/L) requiring replacement, vs. none in the placebo arm.

Headlines announced “Sodium Bicarbonate Failed to Improve Kidney Outcomes in ICU” — which is technically true, but there’s more to consider.