Jul 232018

Jon-Emile S. Kenny MD [@heart_lung]

What is REAL?”

-Velveteen Rabbit

A 42 year old woman with poorly-controlled type II diabetes is admitted with a severe soft tissue infection of her left lower extremity.  She is hypotensive with altered sensorium and she is noted to have a rapidly progressing border of deep, crimson erythema and edema upwards along her inner thigh; she had been recently scratched by her cat there.  Her serum glucose is over 700 mg/dL and her creatinine has eclipsed 3 times her known baseline.  While her urinalysis is without ketones, she has an anion gap of 22 mEq/L with an albumin of 4.4 mg/dL.  Her lactate is 8.0 mmol/L with an arterial pH of 7.14.  She is noted to have deep respirations with a measured PaCO2 of 22 mmHg and a venous CO2 tension of 38 mmHg drawn from a subclavian central line placed in the ED.  The evening critical care fellow opts to administer sodium bicarbonate, but the admitting ICU resident asks her why sodium bicarbonate is indicated in the setting of ‘lactic acidosis.’  


Severe metabolic acidemia – defined here as an arterial pH of < 7.21 with an arterial carbon dioxide tension [PaCO2] of < 46 mmHg and serum bicarbonate less than 21 mmol/L – is known to complicate severe critical illness.  When the cause of the metabolic acidemia is a consequence of a clear bicarbonate-wasting process (e.g. gut or renal loss – non-anion gap acidoses) sodium bicarbonate replacement is typically administered without question.  Yet, as an intern in the ICU, I distinctly recall being chastised for suggesting sodium bicarbonate for a septic patient with severe acidemia and an elevated lactate concentration.  I was handed this paper, and asked to present it on rounds the next day – following a 30-hour shift.

Neatly-arranged into clearly-defined clinical questions, the paper by Forsythe and Schmidt unmasked the absence of evidence for treating ‘lactic acidosis with sodium bicarbonate like a vexed Scooby-Doo villain.  I imagined bicarbonate thrusting its fist into the air and exclaiming that he would have gotten away with it had it not been for Schmidt and Forsythe’s cutting review.  In their appraisal, Forsythe and Schmidt described the dissociation between arterial pH and other body compartments in response to bicarbonate therapy as well as evidence that acidemia does not significantly impair hemodynamics or, on that contrary, that bicarbonate infusion mends cardiovascular malfunction.  Further, they highlighted unwanted consequences of sodium bicarbonate therapy including hypertonicity, hypocalcemia and increased carbon dioxide production as the bicarbonate buffers protons in tissue.

With the aforementioned, I was intrigued to read the recent BICAR-ICU trial by Jaber and colleagues – where patients were treated with 4.2% sodium bicarbonate if there was severe metabolic acidosis – not due to bicarbonate loss, ketoacidosis or exogenous acid ingestion.  Notably 78% of the control group and 86% of the bicarbonate group had a lactate level above 2 mmol/L on randomization with mean values of 5.3 and 6.3 mmol/L, respectively.

What They Did

In a number of intensive care units [ICUs] across France, 398 critically-ill adults with severe metabolic acidosis – defined as a pH < 7.20, PaCO2 < 45 mmHg, bicarbonate concentration < 20 mEq/L plus a SOFA score of at least 4 or arterial lactate concentration > 2 mmol/L – were randomized to receive no additional therapy [i.e. the control group] or to receive aliquots of 4.2% sodium bicarbonate.  Notably, patients were included in the trial only if the underlying etiology of the metabolic acidosis was not a clear bicarbonate-wasting process or if ketoacidosis was the culprit.  The goal of bicarbonate infusion was to raise the arterial pH above 7.30 in the first 48 hours following randomization.  Using an intention-to-treat analysis, all patients were followed for the primary, composite end-point of mortality within the first 28 days and at least one organ dysfunction within the first week of randomization.  Patients older than 65, those with sepsis or with acute kidney injury were stratified, a priori, for analysis.

What They Found

Amongst all patients, there was no statistically-significant difference between the control and bicarbonate groups with respect to the primary, composite end-point of 28-day mortality and 7-day organ dysfunction [71% versus 66%, respectively].  There was, however, a significant reduction in the need for renal replacement therapy [RRT] for those who received bicarbonate as compared to the control group [i.e. an absolute risk reduction of 17%].  Interestingly, the cumulative fluid infused into both groups was not different; the bicarbonate group received, on average, 250 mmol of sodium bicarbonate in the first 24 hours, while those in the control group received none.  Only 26% of patients in control reached a pH of 7.30 for the first 36 hours, while 60% of those who received bicarbonate did.  The anion gap of both groups was, on average, 20-22 mEq/L upon enrollment and this appeared to normalize by 48 hours following randomization.  Within the predefined group of patients presenting with acute kidney injury, both mortality and need for RRT was reduced with bicarbonate as compared to control.


No benefit, no harm

Firstly, there was no statistically-significant reduction in the composite end-point of 28-day mortality and 7-day organ dysfunction following randomization; there was a clinically-significant trend in favour of bicarbonate, but this may have been due to chance.  It remains unclear as to why active alkalinization of the blood in the setting of non-bicarbonate wasting metatolic acidoses provided no clear statistically-significant benefit in the primary end-point.  As highlighted by Schmidt and Forsythe, one reason may be the complicated, multi-compartmental nature of human tissue.  As described previously, capillary and venous-outflow pH is determined not only by the arterial pH entering the tissue, but also by the flow rate through said tissue.  That is, as micro-circulatory flow falls, venous effluent pH also falls as more protons are buffered per volume of passing blood.  Accordingly, arterial pH can be perfect, but if flow is impaired, then capillary and, therefore, venous pH will be low.  As mentioned by the accompanying commentary, an interesting variable to measure would have been central venous blood gas – to assess for changes in venous pH.  Additionally, despite concerns of hypocalcemia and hypertonicity, there was no demonstrable harm to patients who received sodium bicarbonate – though patients who received sodium bicarbonate did have higher rates of hypocalcemia and hypernatremia as compared to control.

Renal protection & lives saved?

Intriguingly, in patients presenting with acute kidney injury there was both mortality benefit and diminished need for RRT in those who received sodium bicarbonate.  As well, RRT was reduced in all patients who received bicarbonate.  Why might this have been?  In the supplementary material it is shown that at 48 hours, arterial pH between the two groups, on average, was strikingly similar.  This may have been due to some cross-over – 47 control patients received bicarbonate as ‘salvage therapy’ after randomization – but more likely, a large proportion of control patients received RRT by 48 hours.  In the trial, RRT was recommended when two of three criteria were present: urine output less than 0.3 mL/kg per h for at least 24 h, arterial pH less than 7.20 despite resuscitation, and kalaemia more than 6.5 mmol/L.  It was the latter two criteria that seemed to drive RRT initiation in the control group as compared to the bicarbonate group.  As seen in their figure 3, the rate of RRT appeared to be at least double in the control group as compared to the bicarbonate group.

Why is mortality lower in those with AKI who received bicarbonate?  One provocative hypothesis is that early, aggressive RRT raises mortality – from complications inherent to urgent RRT – and that delaying early RRT with bicarbonate may prevent this.  For example, in the AKIKI trial, there was increased risk of catheter-related blood stream infection in the early initiation group; most importantly, however, overall mortality was not different between early and delayed RRT in AKIKI – rendering unlikely a hypothesis of early RRT as dangerous.  Thus, we are left with something inherently beneficial to sodium bicarbonate therapy in patients with severe metabolic acidemia presenting with AKI.  Given the lackluster data on the hemodynamic advantages of bicarbonate, salutary effects of alkalinzation may be prevention of arrhythmia and hyperkalemia as well as improved renal blood flow.  Regardless, the real reason remains illusory and intensivists may now reach for 4.2% sodium bicarbonate with more certainty in severe, acute metabolic acidosis – especially in those with acute kidney injury.

Return to the Case

The ICU fellow explains to the resident that the patient’s anion gap acidosis is a consequence of both lactate and uremia.  She also explains that this patient meets inclusion criteria for the BICAR-ICU trial - even with this degree of lactate elevation.  Further, she explained that BICAR-ICU provides evidence that initiating bicarbonate may lower the patient’s risk of kidney injury requiring RRT.



Dr. Kenny is the cofounder and Chief Medical Officer of Flosonics Medical; he is also the creator and author of a free hemodynamic curriculum at heart-lung.org

Get our weekly email update, and explore our library of practice updates and review articles.

PulmCCM is an independent publication not affiliated with or endorsed by any organization, society or journal referenced on the website. (Terms of Use | Privacy Policy)


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