Response to 100mmol of sodium bicarbonate

These are the physiological effects of infusing 100mmol of concentrated (8.4%) sodium bicarbonate into a patient.

What am I giving?

A 1 molar solution of sodium bicarbonate is what you are giving. The osmolality is 2000mosm/L.

Changes to the initial conditions

Let us unfocus from the movements of water and sodium, as they are predictable, and their patterns already well rehearsed. Let us instead observe the traffic of the HCO₃^- anion.

Let us pretend that suddenly 100mmol of this anion is dumped into the extracellular fluid (and being easily water soluble, it frolics merrily through the extracellular fluid compartments, distributing evenly among them). This means 25mmol of HCO₃^- is now in the vascular compartment and 75 mmol is in the interstitial fluid.

The extracellular concentration of bicarbonate pre-infusion in our model is 24mmol/L, which gives us 336 mmol overall. A sudden increase by 100mmol (to a total content of 436 mmol) would cause the concentration to rise to 31.1mmol/L.

Is this what actually happens?

The change in extracellular bicarbonate concentration following a bicarbonate infusion

From the above calculations, it would seem that the volume of distribution for bicarbonate is the same as the extracellular fluid, 14L or about 0.2L/Kg. Experimental findings demonstrate that this is not the case.

Simplistic fluid-filled cylinder models of distribution do not do justice to the complexity of bicarbonate distribution. The major source of complexity, is the tendency of the bicarbonate to buffer hydrogen ions and become "lost" in the process, converting to water and carbon dioxide. This tendency, as one might imagine, is dependent on the presence of acidosis or alkalosis.

An excellent article has examined this relationship, and I will clumsily paraphrase Figure 5 from it below. The figure describes the relationship of serum bicarbonate to its apparent volume of distribution.

In essence, the volume of distribution of bicarbonate changes according to the pH; or rather, according to the pre-infusion bicarbonate level. It is about 50% of body weight under normal conditions (about double the volume of extracellular fluid, just short of the volume of total body water).

At very low pH much of the infused bicarbonate is used up as a buffer, and thus the change in bicarbonate concentration is small (which gives a large apparent volume of distribution, equivalent to 100% of body weight or even more). At high bicarbonate concentrations, more infused bicarbonate remains in the body fluid solution (as there is less acid to buffer) and therefore the volume of distribution remains small. Fernandez et al demonstrate that at high bicarbonate concentrations, its volume of distribution asymptotically approaches 0.4L/kg.

Thus; in our simplistic 42-litre model of a 70kg human, the bicarbonate distributes into 35 litres, with the resulting extracellular bicarbonate concentration of 26.9mmol/L.

Changes in fluid distribution among compartments in an idealised fluid model

For the purposes of calculating osmolar shifts, one can calculate that the net gain of bicarbonate anion for the extracellular fluid is not 100mmol, but 40mmol. The rest vanishes mysteriously into the shaded corridors of buffer chemistry.

Let us pretend that the compartments retain all this extra bicarbonate and all this sodium without trying to excrete or metabolise anything.

Firstly, the increased extracellular tonicity would drag water out into the extracellular space, diluting the bicarbonate concentration. Nevermind the HCO₃^- , there is also an excess 100mmol of sodium in the ECF, and combined with the extra 100ml of water this would result in a 500ml increase in volume for the extracellular fluid, of which 125ml is intravascular.

Distributed in this new expanded ECF volume, the 336 + 40 mmol of HCO₃^- gives a concentration of around 25.9 mmol/L. This is still well above the normal 24mmol/L concentration, constituting an alkalosis.

Will the organism allow this?

No, it will not.

Metabolic fate of infused bicarbonate

Some of the excess bicarbonate will be excreted rapidly by the kidney, which has a massive capacity for it.

(or rather, the kidney has a massive capacity to reabsorb the freely filtered bicarbonate, and it lets some bicarbonate slip out into the urine once a certain concentration threshold is reached)

What is that threshold? Well. For some reason, it has been extensively studied among human infants and anaesthetised dogs. It seems that at a bicarbonate concentration of around 24-26mmol/L, the renal reabsorption of bicarbonate is at its peak. In other words, the peak rate of resorption is at a rate of about 26 mmol per litre of glomerular filtrate. This means, if the glomerular filtrate contains more bicarbonate - lets say 30mmol/L- the extra 4 mmol/L are excreted in the urine without being reabsorbed.

The kidney shares responsibility for the disappearing bicarbonate with the respiratory excretion.

Bicarbonate, after all, is part of a buffering system, which resists changes to the acid-base equilibrium..

It is one simple proton association away from being carbonic acid, which is converted rapidly into CO₂ and water. An excess of bicarbonate thus absorbs some of the hydrogen ions and results in an excess of carbonic acid, which is converted into CO₂ and ventilated away.

The question is, how much of the bicarbonate will be renally excreted, how much will be exhaled, and how much will remain?

This will depend on the presence of other buffers. Bicarbonate accounts for around 80% of extracellular buffering - the rest is hemoglobin (15%) and plasma proteins (5%). These other buffer systems can donate protons to the solution, which bind to even more bicarbonate and thus produce even more CO₂ than would otherwise be possible.

Experimental findings in patients receiving bicarbonate infusions

Studies have been conducted in ventilated patients receiving a load of bicarbonate. They confirm that the amount of infused bicarbonate which is converted to CO₂ is dependent on other extracellular buffering systems.

In the specific study I quote, a 1.5mmol/kg solution of bicarbonate was infused over 5 minutes into a series of ventilated ICU patients (making about 105mmol per patient, if we assume they were 70kg each). This infusion was accompanied by a rise in PaCO2 peaking at the end of the infusion, and then rapidly returning to baseline with 30 minutes. I paraphrase the findings of this study here and produce a stylised graph which resembles the published paper, but I do so without permission of the authors, whom I have not attempted to contact.

The study found that the arterial CO₂ concentration peaks about 1 minute after the end of the infusion, and quickly returns to baseline. The change was significant - from 40mmHg at baseline to 46mmHg at peak.

What is more, this change was greatest in patients with a high serum hemoglobin and albumin, confirming the hypothesis that protein buffers release hydrogen ions to bind with the bicarbonate, resulting in more CO₂ production than would be otherwise possible.

The amount of free hydrogen available at the beginning of a bicarbonate infusion, even in a severely acidotic patient, is miniscule ( 100nmol/L). Therefore only a very small amount of CO2 is produced by the direct effect of infused bicarbonate coming into contact with an acidic body fluid.

However, the act of binding all these protons alkalinises the body fluid. In response, the protein buffer systems release an unknown amount of hydrogen ions, which are in turn bound by the bicarbonate.

This process continues until an equilibrium is reached. This equilibrium will be at a slightly higher pH, with a slightly higher serum bicarbonate concentration - no higher than 26-27 mmol/L, because the kidneys will excrete the excess.

The change in electrolytes following a bicarbonate infusion

Thus, after 30 minutes, a lot of the bicarbonate has been either converted into CO₂ or excreted renally, and the remaining concentration (in a healthy person) cannot be above 26mmol/L, suggesting that in a 14L ECF compartment we have contributed only about 28mmol of HCO₃^-. So really, we have just given a sodium ion load.

The sodium remains extracellular; the water distributes itself accordingly, dehydrating the intracellular compartment by 235ml. The intravascular fluid swells by 83ml and the interstitial fluid gains 251ml. The sodium load raises the sodium concentration from a baseline 140mmol/L to 143.7 mmol/L.

The serum osmolality increases by 2.4mOsm/L, which is slightly less than the 1% osmoreceptor threshold. Similarly, the volume expansion (by 1.6%) does not trigger a baroreceptor response.

So... in summary... what happened to the bicarbonate?

Below is a graph illustrating what might happen to the 100mmols of bicarbonate you infused into a healthy person.

In short, because the patient was healthy, with well-equipped homeostatic mechanisms, the serum bicarbonate did not rise significantly, and the original bicarbonate concentration was rapidly restored by two main mechanisms, renal and respiratory excretion.

What about all this worsening of intracellular acidosis I have been hearing about?

There has been a lot of talk about the "intracellular acidosis" which results from the adminstration of exogenous bicarbonate. As it is highly polar, the extracellular HCO₃^- must be converted to CO₂ and H₂O before it can pass into cells. Carbonic anhydrase effortlessly performs this conversion at an astonishing rate, being among the fastest-working enzymes in the human body.

Thus, CO₂ rather than bicarbonate penetrates into the cells.

The argument has been that this effect serves to paradoxically acidify the intracellular environment. The CO₂, it is argued, once inside the cell becomes carbonic acid again, and dissociates to add protons into the cytosol.

This was based on a study conducted in human platelets, which were exposed to a rapid bicarbonate bolus. The intracellular pH was seen to drop precipitously. The authors concluded that "such treatment is both illogical and dangerous."

In response, several other studies had come to the aid of bicarbonate, demonstrating in rat liver and human leucocytes that bicarbonate boluses actually increase intracellular pH. More studies still have shown that intracellular pH doesnt change. Other still have demonstrated that intracellular pH depends on which buffer solution is used to wash the cells. A confusion has settled over this field.

None of this helps, of course, and there are today as many people who mistrust bicarbonate as there are people who feel it has its rightful uses. This animosity towards it is not helped by the fact that it has been repeatedly shown to increase mortality in the critically ill population. Nor have we been able to demonstrate its usefullness in its traditional roles. For instance, it has not been shown to have any effect on hemodynamic stability or the response to vasopressors in patients with lactic acidosis. In fact, it seems to stimulate the production of lactate, which is counterproductive.

Combined with the above, bicarbonate use is frought with such complications as hyperosomolar hypernatremia, hypocalcemia, volume overload, and cardiovascular depression due to intracellular myocyte alkalosis.

At this stage, if pressed with the question "when would you give sodium bicarbonate?", the fellowship exam candidate would be forced to mention only hyperchloraemic normal anion gap metabolic acidosis, with significant acidaemia.

So which bicarbonate value do I trust?

One has a choice between the bicarbonate in the ABG (which is derived from the pH and the PaCO2 using the Henderson-Hasselbalch equation) and the one in the EUC, which is measured directly. There is only one value, to be precise. The H-H equation describes this value perfectly. And indeed, the two ways of measuring the bicarbonate value agree within 3mmol 96% of the time.