Sodium Bicarbonate Does Not Work in Cardiac Arrest

One of the main arguments for acidosis being harmful is that it is thought to decrease cardiac output. This has been born out in several in vitro studies that showed decreased contractility of isolated cardiac myocytes in an acidemic environment (1,2). However, the clinical application of this data is suspect. As mentioned, these are in vitro studies examining isolated cells of an assortment of animals including rats, rabbits, turtles, guinea pigs, and even trout. The reductions in contractility are minor – around 25% – and mostly occur at an extremely low pH: 6.5 to 6.8 (1,2).

Another proposed reason for acidosis being harmful is that it diminishes response to catecholamines, both among cardiac myocytes and the systemic vasculature, which theoretically could contribute to decreased cardiac output and hypotension (3,4). However, studies again found that most depressed function comes at an extremely low pH, and in vivo, the effect of decreased catecholamine responsiveness is more or less offset by increased catecholamine production during acidemic states (3,4).

So in the laboratory, extreme acidosis probably causes small reductions in cardiac myocyte contractility. But what about in living human beings?

There are two places where the effects of acidosis have been studied in humans. One is in exercise physiology, where lactic acidosis is a natural response to exercise. The other is in patients with acute respiratory distress syndrome (ARDS), where a lung-protective strategy of mechanical ventilation is used. The low tidal volumes associated with this lung-protective strategy result in a permissive hypercapnia, and thus a permissive respiratory acidosis.

Exercise Physiology

Studies measuring pH during exercise on a stationary cycle show ranges as low as 7.1 with a lactate >12.0 mmol/L in healthy subjects. Evaluation of skeletal muscle at low pH shows there is some decreased contractility; however, other physiologic factors likely play a greater role in this decreased contractility than the acidosis itself (5,6). It would reason that exercise-induced acidosis should not cause dysfunction of cardiac and skeletal muscle, as this would make it really hard for humans to survive. Why would we develop an evolutionary response whereupon initiation of physical activity our body enters a dysfunctional state in which our heart and muscles stop working? It’s actually the opposite. Acidosis appears to be beneficial during exertion. Look at the Bohr effect, where lower pH shifts the hemoglobin binding curve, and red blood cells offload more O2 for hypoxic tissues. In this situation, acidosis is beneficial, not harmful.

Acute Respiratory Distress Syndrome

In the ARDS trials, the use of a lung-protective ventilation strategy has likely led to improved outcomes (7,8). This strategy of using lower tidal volumes in mechanically ventilated patients imposes on them a respiratory acidosis, due to the permissive hypercapnia associated with low tidal volumes. The mortality benefit of the lung-protective ventilation strategy persists despite the resulting low pH. It should be noted that in the ICU, acidosis is often buffered with sodium bicarbonate infusion when pH drops very low, i.e below 7.15, although it is unclear if this changes outcomes (9).

In summary, the basis for the theory that acidosis is bad for you comes primarily from in vitro experiments on isolated cells at a very low pH. On the other hand, evidence in living humans suggests mild to moderate acidosis does not have deleterious effects and is potentially beneficial. So why would we correct acidosis at all? Well, as mentioned above, the decreased contractility of cardiac myocytes was only seen at severely low pH, so perhaps this is where the bicarbonate infusion will help. Exercise-induced acidosis was not harmful in healthy patients, but can we generalize this finding to the typical ED or ICU patient? And in our ICU patients in whom permissive hypercapnia was well tolerated, perhaps there was such a substantial mortality benefit from using low tidal volumes that it masked any negative effects of the resulting respiratory acidosis. It is hard to extrapolate evidence from one population to another.

Infusion of sodium bicarbonate – NaHCO3 – works by directly buffering H+ ions. HCO3 combines with H+ to form H2CO3 – carbonic acid – which then splits into H2O and CO2. The CO2 is ventilated off. Therefore, bicarbonate really functions by changing your metabolic acidosis to respiratory acidosis. As long as you can adequately ventilate off the CO2 that is generated by bicarbonate, your acidosis will improve. Conversely, if a patient already has functional respiratory acidosis, the addition of more CO2 to the picture is not going to help them and will actually make their acidosis worse. So, adequate ventilation is 100% necessary for sodium bicarbonate to work. 

There are a few inherent negative effects of sodium bicarbonate therapy. Studies have found rapid bicarbonate administration will cause a paradoxical worsening of intracellular acidosis despite alkalinization of the extracellular fluid (10). This is from an imbalance in CO2 across cell membranes. As mentioned, bicarbonate administration will take the H+ in your blood and convert it to H2CO3, which then quickly breaks down to H2O and CO2. So there is a time there where your extracellular CO2 rapidly rises before you can ventilate it off. That rapid increase in CO2 diffuses almost instantly across cell membranes, where the reverse reaction occurs: CO2 + H2O leads to increased H+ in the cell thus generating an intracellular acidosis.

Sodium bicarbonate also includes a huge load of sodium which can create hypernatremia and hyperosmolality, expanding the extracellular circulating volume (ECV) quickly. One dose of the sodium bicarbonate found in crash carts – 50 mL or 50 meq of 8.4% sodium bicarbonate – will raise the Na by 1.0 meq and increase ECV by 250 mL.

The theory behind using sodium bicarbonate in cardiac arrest seems to be that it may counteract the severe lactic acidosis resulting from global hypoperfusion. Remember that severe acidosis may cause decreased responsiveness to catecholamines – i.e. your epinephrine pushes won’t work – and an increased risk of dysrhythmias – i.e. that VF won’t quit. However, these theoretical harms of acidosis are primarily based on animal and in vitro studies. When you look at the human trials studying the use of sodium bicarbonate in cardiac arrest, it pretty clearly shows no benefit and even possible harm (11). To date, there are only two RCTs studying sodium bicarbonate in cardiac arrest, and both showed no benefit (12,13).

The first trial was a randomized, double-blind, placebo-controlled study of 502 patients 16 years or older with out-of-hospital cardiac arrest – asystole or VF refractory to the first defibrillation (12). They found no difference in outcomes between patients who did and did not receive buffer therapy.

The second trial was also a randomized, double-blind, placebo-controlled study of out-of-hospital cardiac arrest – this time with 874 patients (13). They studied a similar cohort of patients, including anyone over 18 years old with cardiac arrest refractory to defibrillation. They found no difference in their primary outcome between patients who did and did not receive bicarbonate.

There are some issues with this study (13). First, the primary outcome of the trial was having a pulse on ED arrival, which is not a patient-centered outcome. In the trial, about 14% of patients achieved this primary outcome. But if you look at the data from around the world, the percentages of patients with meaningful survival after out of hospital cardiac arrest are around 7-9% (14,15). So ⅓ or ½ of the patients with a supposed good outcome in this study likely did not go on to meaningful survival. And if the primary outcome is not meaningful, it’s hard to know how to interpret the rest of the data. 

The second problem is the headline finding in this study (13). Vukmir et al report that patients with prolonged cardiac arrest had increased survival if they received bicarbonate therapy. They found this outcome in a subgroup analysis of patients with cardiac arrest for 15 minutes before CPR initiation. However, on close inspection, this is misleading. 

Vukmir et al report that giving sodium bicarbonate to patients who had prolonged downtime doubled their chance of survival from 15% to 33% (13). Those numbers alone should raise suspicion. They are saying that in one cohort, 33% of patients who did not receive CPR for over 15 minutes survived. That’s better than their overall survival rate of 14%. So according to their conclusion, it was beneficial to have been dead for more than 15 minutes before CPR – it doubled the chance of survival. In comparison, although it reported a longer-term outcome, a study of 30,000 patients with out of hospital cardiac arrest found that among patients in whom CPR was initiated after 15 minutes, 30-day survival was 0.9% (16).

Obviously, something is not right here. It turns out Vukmir et al misinterpreted the numbers. What they did: they looked at all the patients who survived with prolonged downtime, and saw twice as many received bicarbonate (red box to the right). They then incorrectly concluded that giving bicarbonate to patients with prolonged downtime doubled the chance of survival. But this doesn’t make sense, because everyone in that analysis survived. How can you say bicarbonate doubled the chance of survival if nobody died? The correct interpretation of these numbers is: patients with prolonged downtime were twice as likely to receive bicarbonate therapy.

To see if there actually was a survival benefit in this subgroup, you will have to look at what happened to the patients who did not survive. We re-analyzed the data, taking the patients with prolonged downtime and comparing survival vs no survival with bicarb vs no bicarb (red box to the left). Using the Fisher’s exact test, we found a p-value of 0.1 (not statistically significant) – meaning there is no difference in survival in patients with prolonged arrest. So, aside from the issue of titling a paper with the result of a subgroup analysis, the main conclusion of the paper is inaccurate.

Besides these two RCTs (12,13), there are numerous retrospective, observational, and cohort studies of sodium bicarbonate administration in cardiac arrest that show no benefit or harm (11). In fact, the 2015 AHA guidelines state “sodium bicarbonate should not be used routinely in cardiac arrest” unless you suspect arrest from TCA overdose or hyperkalemia (17).

1. Acidosis itself is likely not as harmful as we think.
2. Sodium bicarbonate therapy works by changing a metabolic acidosis to a respiratory acidosis, so you must be able to ventilate off the extra CO2 to raise the pH.
3. There are several negative effects of sodium bicarbonate therapy, including worsening intracellular acidosis and increased extracellular volume.
4. The literature behind using sodium bicarbonate in undifferentiated cardiac arrest clearly shows it does not work and may even be harmful. The AHA recommends against its routine use. So stop using it.