“I should refresh myself on acid-base stuff” I naively thought to myself, after being asked to give some impromptu teaching at an ALS course. “Probably just recap the bicarb-CO2 thingy and that’ll cover it”.

Oh, sweet summer child.

After many hours of trying to find the ABC answer to this topic, I’ve discovered no such certainty exists. There is no watertight, nicely presented framework or theoretical underpinning to the problem of acid-base balance. Instead, there is a Wild West of chemists slinging contentious theories at each other, with multiple differing algorithms and approaches. It’s a black hole of never-ending tangents and off-shoots.

Disclaimer: I have selectively included and ignored aspects of this treacherous topic at my own whim. For example, I have not covered the Six Simultaneous Equations used by Stewart, nor other such physiochemical wizardry. See references yourself if you want to get into the weeds.

Why is acid-base important?

Probably a reasonable first question. Generally, bodily processes operate within a tight pH between 7.35 and 7.45. 

  • In chemistry, neutral pH is 7. The body tends to be slightly more alkali (around 7.4) because some intermediates involved in biochemical reactions can become ionised at a pH of 7.  
  • pH affects the O2-Hb dissociation curve. Oxygen offloads from Hb more readily in acidic environments, such as metabolic tissues.
  • Protein structure and function becomes more unstable and denatured in lower pH.
  • Acidosis is negatively inotropic, initially compensated by catecholamine drive, but worsening with pH <7.2. Risk of tachycardia and arrhythmia is also higher in acidosis. 

There are two key questions before even looking at a blood gas:

  1. Do you need one?
    • A blood gas will tell us many things, but uniquely it can give insight into 1) ventilation/oxygenation and 2) acid-base balance. I will be primarily focusing on acid-base balance in this article.
    • Thankfully the fetish for ABGs in every patient with even mild hypoxia is becoming less prevalent, but always consider how a blood gas will influence management. Patient just had a tonic-clonic seizure in front of you? Don’t get the gas for a lactate (it will be high), but it may be useful to see their sodium. I’ve heard of clinically well patients being kept on wards just because their lactate is slightly above normal on repeat gases. That leads us to…
  2. How does the patient look?
    • You need to know the clinical context attached to the numbers. A mild acidosis may be no big deal in a diabetic ketoacidosis, but could be extremely worrying in a calcium channel blocker overdose. A study of 6 elite rowers found an average lactate of 26 and pH of 6.85 after 2000 metres rowing; needless to say they didn’t need intubated and wheeled off to critical care. pH on its own doesn’t matter; the pathology driving the numbers is what counts. 

Bare bones chemistry

I’m an emergency physician, I need things to be as simple as possible.

What does acid-base balance actually mean? Essentially, its the interaction between different systems within the body to influence hydrogen ion concentration and, thus, pH.

Hydrogen is an atom.

A hydrogen ion is a hydrogen atom without its orbital electron; functionally, therefore, it exists as a proton. It is depicted as H+ due to only being a hydrogen ion with a single proton, with the two terms often being used interchangeably. 

Hydrogen ions exist in much smaller concentrations in the blood than other ions, and so are measured in nanomoles (thousand-millionths of a mole). For comparison, sodium, potassium and other electrolytes are measured in millimoles (thousandths of a mole).

The absolute number of hydrogen ions may be way smaller than other ions, but we tolerate a higher margin of relative flux in hydrogen ion concentration. For example, a 2.5x increase in [H+] from 40 to 100 nmol (equivalent change of pH 7.4 to 7.0) is survivable, but a 2.5x flux in sodium or potassium would kills us.

From Chawla et al (2008)

Fundamentally, acidosis is a rise in hydrogen ion concentration [H+] whilst alkalosis is a reduction in [H+]. 

Hydrogen ions do not generally exist in the body in freely, dissociated form as protons just floating around. If not buffered by substances such as proteins, Hb or phosphate, they are associated with water as H₃O⁺. 

H+ + H2​O → H3​O+

For simplicity, this form of associated hydrogen ion is usually just referred to as ‘free’ hydrogen ions. 

pH is a representation of the concentration of hydrogen ions ([H+]), but is not in itself a concentration. Because of this, pH does not have any units: it is just a number. Hydrogen ion concentration determines the pH and, therefore, the acidity of a solution. 


Henderson-Hasselbalch

The Henderson-Hasselbalch (H-H) equation gave us a shorthand for measuring the acidity or alkalinity of a solution; pH. This provides a measurable and sensible representation of hydrogen ion concentration.

  • pH = a measure of the acidity/alkalinity. Expressed as the negative logarithm of [H+] – ie a lower pH means a higher [H+]
  • pKa = how strong an acid is, determined by how easily it dissociates to H+ and its conjugate base in solution
  • [A-] = concentration of conjugate base, after HA has dissociated
  • [HA] = concentration of acid that hasn’t yet dissociated and given up its H+
  • Log = to keep measurements of [H+] within a reasonable, understandable range, the logarithmic scale is used. This is a nonlinear scale, where a change of one pH unit reflects a 10-fold change in hydrogen ion concentration [H+]. A normal blood pH of 7.4 reflects a [H+] of 40 nmol/l, whereas at pH 7.0 it is 100 nmol/l. 

Using the above equation, one can work out, say, the pH of hydrochloric acid (HCl). 

So, what actually causes a rise or fall in pH in the body?

Two Broad Schools:

The first thing to say is: nobody fully knows what causes fluctuations in acid-base balance in the body. But two broad schools of thought have evolved as a framework to explain what may be going on:


1) Traditional Approach

Building on work by Arrhenius in the 1800’s, Bronsted and Lowry established our traditional clinical understanding of acid-base chemistry in 1923. The key idea they proposed is that acids donate hydrogen ions, whereas bases accept them. 

  • An acid donates H+; after donating, it becomes a conjugate base, due to loss of a proton.
  • A base accepts H+; after accepting, it becomes a conjugate acid, due to gain of a proton.

For example:

HCl + H₂O → Cl⁻ + H₃O⁺

  • HCl is an acid, that donates a proton to H2O, to become a conjugate base (Cl-)
  • H2O is a base, that accepts a proton, to become a conjugate acid (H₃O⁺)
  • The generation of H₃O⁺ results in a reduction of pH through increased acidity.
  • If we were to add HCl to water, this increases acidity (reduces the pH) due to the binding of H+ to neutral water molecules.

So the traditional approach holds that if we add HCl to a solution, pH will reduce and acidity will increase as a direct result of the addition of H+. 

As this theory became ingrained in clinical practice, attention was increasingly focused on PaCO2 and HCO3- as the primary determinants of bodily pH. This was further cemented by incorporating carbonic acid an bicarbonate into the Henderson-Hasselbach equation.


The carbonic acid equilibrium equation

CO2 + H2O <-> H2CO3 <-> HCO3– + H+

This chemical equation underpins most of clinical reasoning in acid-base interpretation and does a lot of heavy lifting in the interpretation of acid-base disturbance.

It placed CO2, H+ and bicarbonate at the centre of our clinical understanding of acid-base, leaning on these three as the sole mechanisms for acidosis and alkalosis.

Whilst it is accurate and sufficient for respiratory acid-base interpretations, we should be aware that it only provides a superficial, incomplete explanation about what’s going on metabolically. 

It’s main shortcoming is that it relies on only using CO2, HCO3– and H+ to explain all disturbances in acid-base physiology. Unfortunately, there is a lot more going on metabolically than just these three.

2) Stewart Approach

Peter Stewart challenged this Traditional approach in 1978, asserting that early definitions of acid-base by Arrhenius were more accurate than those of Bronsted and Lowry.

Specifically, instead of acids being seen as proton donors, he defined an acid as any ion that shifts the dissociation equilibrium of water to higher concentrations of hydrogen ions.

H2​O ⇌ H+ + OH−

Rather than an acid giving up a proton (H+), he contested that an acid would directly cause a shift in our bodily water, where H+ would be liberated from us rather than added to us from an external source.

He asserted that pH depends on 3 laws:

  • Law of conservation of mass
  • Law of mass action
  • Law of electroneutrality 

For my own sanity and simplicity, I will focus on the third law as the most explainable and relevant. 

Whereas Bronsted-Lowry theory asserts that pH is affected directly by the addition or subtraction of hydrogen ions in a solution, Stewart proposed it is the electrochemical charge of ions that ultimately alters pH.  

Simply put, the law of electroneutrality dictates that the electrochemical charge of a solution in the body must remain neutral. If this is upset, then hydrogen ions are liberated from, or formed into, water, to increase or decrease net electircal charge in a compensatory response. 

Using the same example of HCl as above, when dissociated, Cl⁻ increases the net negative charge in the body as it is a negatively charged ion. A lack of adaptive response would violate the law of electroneutrality. 

Therefore, in response to this increase in net negative charge from chloride ions, water within the body dissociates to release hydrogen ions (which are positively charged), to restore electrochemical balance to a neutral charge. 

H2​O ⇌ H+ + OH−

Stewart asserted that this water dissociation equilibrium above provides an inexhaustible source or sink for hydrogen ions. 

Q: Why doesn’t the H+ from HCl directly contribute to acidosis if it is agreed that H+ is responsible for a change in pH?

The H+ that results from HCl is immediately buffered by any number of buffering systems:

  • Bicarbonate: H+ combines with HCO3- to eventually be ventilated off as CO2
  • Intracellular buffers: protein (mostly Hb) and phosphate
  • Serum buffers: albumin and phosphate
  • Ammonium and urinary buffers

So the H+ that is released from HCl is rapidly buffered by the body. This leaves Cl- as a strong ion unbuffered. In this case, H+ has to be liberated from body water. The initial H+ from HCl is unable to contribute to electroneutrality as it has been buffered already, eg ventilated off as CO2.

Unlike Bronsted-Lowry, Stewart believed that directly adding or removing hydrogen ions would have no effect on the pH, as this water dissociation equilibrium would accordingly shift left or right to maintain the net concentration of hydrogen ions (and thus electroneutrality), thereby maintaining neutral pH. 

In contrast, adding a strong negative ion such as Cl⁻ from HCl will result in an increased liberation of positively charged hydrogen ions from water in the body, in order to balance the net negative charge from Cl⁻. This is because Cl⁻ cannot be directly ‘neutralised’, and so hydrogen ions are liberated in response.

In a different type of example, sodium bicarbonate increases alkalosis not because of the addition of bicarbonate, but because sodium cations increase the positive net charge in the body. In response to this, H+ concentration reduces to maintain electroneutrality, and thus pH increases.

Q: What are strong acids, weak acids, and other non-electrolytes?

  • Strong acids and bases in solution are always fully dissociated into their constituent ions. For example, sodium chloride (NaCl) is always dissociated into Na+ and Cl- in solution. Most strong ions are inorganic (eg Na+, Cl-, K+) but some are organic (eg lactate).
  • Weak acids are those that are only partially dissociated into its constituent ions. HA ⇌ H⁺ + A⁻ represents this equilibrium, where HA is the acid, and A⁻ is its conjugate base. A weak acid is in dynamic equilibrium between its undissociated (HA) and dissociated (H⁺ + A⁻) forms. Carbonic acid is an example of a weak acid, mediating the reaction between carbon dioxide + water, and bicarbonate and hydrogen ions. 
  • Non-electrolytes never dissociate into ions, so only contribute to the osmolality but not the charge in a solution. 
In the traditional approach, acidosis results as a direct effect of adding H+ to a solution. The Stewart/Physiochemical approach: HCl dissociates into H+ and Cl- when added to solution. H+ is immediately buffered (by Hb, proteins, phosphate, albumin, ventilated off etc). The remaining Cl- reduces net charge of the plasma, and so H+ ions are liberated from body water. This increased H+ concentration reduces pH.

Q: What is the role of OH- in all this? Doesn’t its existence just balance out any H+ added?

Firstly, water can dissociate into H+ and OH-, but its absolute amount of these ions is actually very low.

Second, water does not exist in a 1:1 relationship with H+ and OH-.

The H+ and OH- ions from water exist in unequal amounts in the body, due to their involvement in other processes. For example, some of the H+ from H2O dissociation will be immediately buffered in order to maintain the slightly alkaline pH of 7.4, meaning that there is generally a larger amount of OH- than H+.

Although there may be proportionately more OH- than H+ to maintain neutral body pH, the addition of a strong ion or weak acid will reduce pH by liberating more H+ to maintain electroneutrality.

Different acid and base physiologic processes in the body will also mean that H+ and OH- are in slightly varying amounts between each other.

What happens to the OH- from H2O dissociation in response to the addition of HCl? I can’t get a satisfying answer to this, and probably draw my current limit for further digging here. All I need to know is that the ↑ H+ results in acidosis, and OH- goes ‘somewhere’.

Q: Is water a type of buffer then?

Water is not classed as a buffering system in the body for the following reasons:

  • The reaction between H+ and OH- is simply a neutralisation reaction, which occurs once but does not buffer against further changes.
  • Water, once neutralized, no longer has OH⁻ available to react with additional H⁺. It has a very low capacity to neutralise H+ due to the low amounts of OH-. Water’s dissociation is very limited, and it doesn’t have a reservoir of ions to buffer changes in pH after initial neutralisation.
  • Water does not work in a dynamic equilibrium. Unlike carbonic acid, water has a very limited ability to flex back and forth between its ionic constituents and H2O. This is because it is not a buffer system that can function over a range of pH changes, but exists as a less dynamic equilibrium reaction.

A quick checkpoint:

After that initial foray into acid-base history, its useful to pause and take stock.

For my own sake, here is a brief summary of conclusions we can draw so far.

  1. Serum hydrogen ion (H+) concentration changes pH and acid-base status.
  2. It is likely that electrochemical imbalance, rather than direct addition of H+ or HCO3- per se, is the main driver of metabolic acid-base disturbance.
  3. The carbonic acid equilibrium equation is a good explanation and clinically useful for respiratory acid-base disturbance. This holds true across both the Traditional and Stewart approaches.

IMPORTANT CAVEAT:
Stewart was a science bro, not a clinician. This means that although his approach may provide a deeper understanding of what causes acid-base disturbance, trying to wholesale adopt his entire framework will usually be prohibitively unwieldy at the bedside.

This doesn’t stop some clinicians doing their darndest to turn bedside acid-base analysis into a math degree.

Nonetheless, after gaining an appreciation for this, we can parse away and incorporate that which is helpful.

Respiratory Acid-Base balance

Mercifully, there is broad agreement between the Traditional and Stewart approach when it comes to explaining respiratory acid-base balance.

We don’t need to throw the carbonic acid baby out with the protonated bathwater, and can thankfully still cling onto this much touted pillar of acid-base teaching; at least as far as respiratory perturbations are concerned.

Carbon Dioxide / pCO2

At rest, an adult produces about 200-250mls of CO2 per minute, which is ventilated off via the lungs, maintaining arterial pCO2 within a normal physiologic range. 

CO2 + H2O <-> H2CO3 <-> HCO3– + H+

A respiratory acidosis occurs when there is a rise in the partial pressure of CO2 in the blood. This can occur from increased presence of inspired CO2 (not common), increased production of CO2 in the body (such as malignant hyperthermia and laparoscopic CO2), or decreased alveolar ventilation (nearly all cases in clinical practice are due to this). A change in the partial pressure of CO2 will influence serum pH. If CO2 rises, then we see a rightward shift in the carbonic acid dissociation equation, resulting in the formation of H+ and HCO3-. This increase in H+ concentration causes a reduced pH and increased acidity. 

During acute respiratory acidosis, H+ is buffered by Hb (shown here) as well as phosphate and other intracellular proteins.

Well why doesn’t the HCO3- produced just buffer this increased H+ concentration? 

The answer is that a system cannot just buffer itself. After creating the H+ and HCO3 from CO2 in our body, we can’t just immediately reverse the reaction back to CO2 and H2O, that wouldn’t make sense. If a respiratory acidosis persists, then the kidneys will increase reabsorption of HCO3-, but this is something that takes 6-12 hours to occur, reaching maximal effect by around 3-4 days.

So if it’s not HCO3- that buffers H+ from an acute respiratory acidosis, then what does?

In the immediate term of a respiratory acidosis, the excess hydrogen concentration is buffered primarily intracellularly by haemoglobin, phosphates and other intracellular proteins. This buffering system can become saturated quite quickly, leaving increased ventilation as the only remaining way of acutely resolving the acidosis. If ventilation doesn’t alter accordingly, hypercapnic respiratory failure and worsening acidosis ensues.

Increased pCO2 has a number of physiologic effects:

  • Rightward shift of the O2-Hb curve; causing reduced Hb affinity for O2 to offload for tissues
  • Cerebral and peripheral vasodilation
  • Pulmonary vasoconstriction; to divert blood from areas of V/Q mismatch to better ventilated areas of lung
  • Sympathetic stimulation
  • Increased ventilation via central (medullary) and peripheral (carotid bodies and aortic arch) chemoreceptors; these detect changes in pH and pCO2. 

The body will respond to a respiratory acidosis rapidly by increasing ventilation; this is because CO2 easily crosses the blood brain barrier to stimulate the respiratory control centre in the medulla. However, at a certain level of CO2, the respiratory drive will be abolished and narcosis will set it; this is what we see with significant respiratory acidosis in, for example, an exacerbation of COPD. 

Metabolic Acid-Base Balance

Thankfully, respiratory acid-base changes are explained simply enough using our well known carbonic acid dissociation equation as above.

Metabolic acid-base disturbance is more complex and contested.

During the Traditional era, two useful concepts were introduced to help address this black-box of metabolic acid-base interpretation:

Base excess and the anion gap.

Base Excess

Base excess is available on a blood gas as a number. The normal is zero, with a physiologic range of -2 to +2. 

What is it?

The number given represents the amount of excess or deficit of base in the blood. A positive base excess implies an excess of base (metabolic alkalosis), whereas if it is negative, this suggests a deficit of base (and a metabolic acidosis, given that base is being used up to buffer protons). 

Base excess (BE) is an aggregate measure of metabolic changes, that considers all buffering capacity; this includes HCO3- and non-bicarbonate buffers (haemoglobin, plasma proteins such as albumin, and phosphates). As such, it is a more sensitive measure of metabolic disturbance than HCO3 alone, as it accounts for all buffer bases, not just HCO3. 

Importantly, BE is unaffected by respiratory acidosis/alkalosis, as it is calculated using a ‘normal’ pCO2 (40 mmHg / 5.3 kPa), and so base excess specifically reflects only metabolic (non-respiratory) changes in bicarbonate and related buffers, but not the immediate respiratory-driven buffering processes. However, a chronic respiratory acidosis would result in raised HCO3-, which would alter the base excess.

Therefore, if PaCO2 is raised but the BE is normal, then there must be an acute respiratory acidosis present without any metabolic acidosis/alkalosis. 

Limitations:

There are some limitations to BE. It only measures blood/serum buffers, and so won’t pick up on the frantic intracellular buffering of H+ from excess CO2 in a respiratory acidosis, for example.

The fact it is an aggregate measure of all buffering systems also means it doesn’t say what specifically is going on, but just points in a direction for metabolic acidosis/alkalosis. 

Because BE is influenced by anions and proteins such as albumin, as well as phosphate, if these are low (like hypoalbuminaemia in liver disease, or hypophosphataemia), the base excess may become more unreliable.

Anion Gap

The anion gap (AG) complements BE by providing a bit more detail about what could be causing a metabolic disturbance. 

AG = ([Na+] + [K+]) − ([Cl−] + [HCO3−])

Anion gap serves to differentiate a non-anion gap metabolic acidosis (NAGMA) from an anion gap metabolic acidosis (AGMA).

  • NAGMA is essentially a hyperchloraemic acidosis
  • AGMA is either:
    • an increase in endogenous acid eg lactate, urea, ketones etc
    • ingestion of exogenous acid eg toxic alcohols, salicylate etc

Anion gap recognises the need for electroneutrality in the body.

As previously explained, this is a fundamental law that states every solution must have a net charge of zero, with an equal balance between all the positive ions (cations) and negative ions (anions). The anion gap is calculated by measuring the difference in charge between the commonly measured extracellular ions; Na+, K+, Cl- and HCO3-. When calculated, this reveals a normal ‘gap’ of around 4-12 mmol/l.

This gap is made up of the unmeasured anions such as lactate, sulfate, ketones, phosphate, and weakly negatively charged proteins such as albumin. This 4-12 mmol/l represents a ‘normal’ amount of this unmeasured stuff in a healthy individual.

If any of these unmeasured anions increases, this will subsequently reduce HCO3- in order to maintain electroneutrality. This reduction in HCO3- will give a raised anion gap value.

Therefore, if there is an acidosis with a widened gap, we can assume that there is an increase in the hidden, unmeasured anions contributing to the metabolic acidosis. This could be ketoacidosis, lactic acidosis, uraemia, salicylate poisining, etc. Its useful to have a mnemonic to sieve through the potential culprits in a raised anion gap acidosis

CAT MUDPILES

  1. Carbon monoxide/cyanide
  2. Aminoglycosides
  3. Theophylline
  4. Methanol
  5. Uremia
  6. Diabetic ketoacidosis
  7. Paracetamol/acetaminophen
  8. Iron/isoniazid
  9. Lactic acidosis
  10. Ethanol/ethylene glycol
  11. Salicylate/ASA

Limitations:

Similar to BE, the anion gap calculation can become less reliable in states of low albumin and phosphate, where a ‘normal’ gap may be hard to define for such individuals. A low albumin may hide a raised anion gap acidosis, by giving a falsely normal/low anion gap. As a rough estimate, every 1 g/L decrease in albumin will decrease the anion gap by 0.25 mmol/L

A correction for albumin in calculating the anion gap can be as follows:

Corrected AG = AG + 0.25 x (40 – albumin (g/L))

This corrected anion gap formula can be used when there are significant deviations from the norm in albumin levels.

A further limitation is that a high anion gap acidosis and normal anion gap acidosis can co-exist, reducing the utility of the anion gap measurement.

Stewart Approach – further refining metabolic acidosis analysis

As stated, the Stewart approach holds that a rise or fall in H+ concentration is dependent on variations in electrochemical charge, rather than the direct addition/subtraction of H+ per se. 

This physiochemical approach assigns H+ and HCO3- as dependent variables that are only altered by other, external and independent variables. So it is not low bicarbonate per se that causes an acidosis, rather it is a reflection and second order effect of another primary causative factor. So HCO3- becomes a marker rather than a mechanism in acid-base analysis. 

The primary dependent variables that can be altered are:

  • H+ (pH)
  • HCO3-

The 3 independent variables that can influence H+ and HCO3- concentration are:

  • Carbon dioxide / pCO2
  • Strong ion difference (SID)
  • Total nonvolatile weak acids (Atot)

We have already established that carbon dioxide / pCO2 is a widely agreed model for respiratory acid-base balance, which leaves us with SID and Atot:

Strong ions and Strong Ion Difference (SID)

Strong ions are those ions that are fully dissociated in solution, both cations (positive ions) and anions (negative ions). 

In the body, these cations are primarily sodium, potassium, magnesium and calcium, whilst the anions include chloride, lactate and ketones. 

So the strong ion (SID) difference is calculated by:

([Na+]+[K+]+[Ca2+]+[Mg2+])−([Cl−]+[lactate−])

Due to the relatively low numbers of the other ions, this equation can be simplified for clinical application to [Na+] – [Cl-]. Potassium, magnesium and calcium are kept in such a tight range that they can clinically be excluded from the calculation. Lactate similarly is normally minimal.

Reference ranges usually for Na (135-145) and Cl (95-108).

A typically normal SID is around 38 to 42 mEq/L. This difference comprises HCO3- and weak acids (plasma proteins (mostly albumin), phosphate and other unmeasured anions).

When calculating a simplified SID, any increase in chloride, or reduction in sodium, will reduce this difference, with anything below 38 representing a hyperchloaremic acidosis. 

This is analogous to a NAGMA (non-anion gap metabolic acidosis).

Wait, how does this rise in anions cause an acidosis?

Remember, a change in net electrical charge from neutral will necessitate an increase/decrease of H+ in as a compensatory response to maintain electroneutrality. This increase in anions, resulting in decreased SID, could be from increased chloride (such as in a hyperchloraemic acidosis) or reduced cations (ie. sodium, such as in water excess). A decrease in SID will result in the following compensatory mechanisms to maintain electroneutrality, resulting in acidosis and reduced pH:

  • Increased H+ liberation from body water 
  • Decreased HCO3- by increased renal excretion of HCO3- and ventilatory excretion of CO2

Bear in mind that the above equation (Na+ – Cl-) will only show an acidosis/alkalosis due to changes in sodium or chloride.

The full SID calculation ([Na+]+[K+]+[Ca2+]+[Mg2+])−([Cl−]+[lactate−]) would reveal an acidosis/alkalosis due to fluctuations in those ions within the equation. 


Total Weak Acids (Atot)

The final independent variable is total weak acids (Atot), comprising mainly albumin and phosphate. In the absence of other anions (such as lactate, ketones etc), Atot should approximately equal the normal anion gap, given this gap is normally comprised of albumin and phosphate (as well as small amounts of other anions).

The amount of serum phosphate is quite small, and so fluctuations rarely produce clinically significant changes in acid-base status. The contribution of albumin to acid-base is significantly larger than phosphate.

Why is albumin a weak acid?

Albumin has many amino acid side chains. Some amino acid side chains (such as those of lysine and arginine) are positively charged. However, the larger amount of carboxyl groups on side chains of other amino acids are negatively charged, as they lose their H+ (-COOH → COO- + H+). This imbalance leaves albumin with a net negative charge (approx 0.25 the negative charge of other strong anions), and it acts as weak acid to liberate H+ from body water to counteract this negative charge. 

Due to this, fluctuations in albumin can influence acid base balance.

Hypoalbuminaemia (often seen in critical care, cirrhosis, etc) will reduce the total amount of weak acid, reducing net negative charge, increasing HCO3- and reducing H+, to give a raised pH with subsequent increased alkalosis.

The reverse would be true for hyperalbuminaemia, such as in administration of albumin solution. ↑ albumin = ↑ weak acid = ↑ negative charge = ↑ H+ liberation = ↓ pH / ↑ acidosis. 

With this in mind, if plasma protein levels are normal, then acid-base disturbances can be solely analysed in terms of changes in the other 2 independent variables; pCO2 and SID. 

When there are significant deviations in normal plasma proteins (mainly albumin), these should be taken into account when analysing acid-base. 


What is the difference between the strong ion difference and anion gap?

The strong ion difference measures independent variables, which are all strong anions. The full equation includes the following:

([Na+]+[K+]+[Ca2+]+[Mg2+])−([Cl−]+[lactate−]). 

Therefore, any change in the measured SID variables will be captured in this equation. These would primarily be hyperchloraemic and lactic acidoses, given acid-base disturbances from fluctuations in the other ions is rare.

On the other hand, the anion gap is measured thus:

Na+ – (Cl- + HCO3-).

The key difference is that the AG includes a dependent variable; HCO3-.

This dependent variable can be influenced by any of the independent variables mentioned earlier (pCO2, SID and Atot). In the context of an acute metabolic acidosis, any increased anion will reduce the measured serum HCO3- in order to maintain electroneutrality. This decrease in HCO3- will result in a raised anion gap, indicating unmeasured anions’ indirect influence on ‘squeezing out’ HCO3- due to their anionic negative charge.

So the inclusion of HCO3- in the anion gap equation allows us to see if there is any increase in the unmeasured anions, by the indirect influence on HCO3- altering the anion gap. Whereas the strong ion difference is only influenced by changes in the ions directly in the equation, given they are independent variables.

My approach:

All this begs the very reasonable question: why overly complicate things?

We have a traditional, workable approach to acid-base interpretation that is easily applied to the vast majority of clinical case. Such ramblings on the topic fly in the face of Occam’s razor, where the simplest explanation is best.

Interpretation of acid-base disorders will always remain partly an art, one that combines an intelligent synthesis of the clinical history, physical examination, and other ancillary laboratory data taken together in the context of the individual patient and the nature and temporal course of his or her disease.” – unknown

The physiochemical/Stewart explanation to acid-base disturbance makes sense to me. It better explains why hyperchloraemia causes an acidosis, the influence of albumin as a weak acid, the role of body water as a sink for H+, and so on.

But I’m not going to apply its numerous equations to the bedside, because it’s absurdly impractical for me.

Trying to understand the theory, whilst clinically applying it in the simplest way possible is my aim.

As deranged as it may be, this is my website, and therefore the following is my own approach to acid-base interpretation:

  1. pH: is there a clear acid-base disturbance?
  2. pCO2: Is there a respiratory contribution?
    • pCO2 and pH go in opposite directions:
      • Resp acidosis: low pH, high pCO2
      • Resp alkalosis: high pH, low pCO2
  3. Base excess & HCO₃⁻: Quick metabolic snapshot
    • BE corrects for any respiratory contribution and provides a pure metabolic impression. Abnormal HCO3- also helps point toward acid-base disturbance.
    • Eg. a bicarb of 10 or BE of -14.0 is useful in highlighting a metabolic acidosis, but neither bicarb nor BE give the cause. 
  4. Simplified SID (Na⁺ – Cl⁻): Sodium/chloride-driven disturbance
    • Quickly ascertain whether this is a sodium/chloride driven disturbance, such as a hyperchloraemic acidosis.
    • Reduced SID (<38) indicates a hyperchloraemic acidosis (may be due to low sodium and/or raised chloride)
  5. Lactate: Check for lactic acidosis
  6. Anion Gap: Check for unmeasured contributors
    • If raised, consider CATMUDPILES
  7. Consider albumin: Adjust anion gap if necessary
    1. If there is hypoalbuminaemia present, correct for this with the following equation:
    2. Corrected AG = AG + 0.25 x (40 – albumin (g/L))

Simplified:

  1. pH
  2. pCO2
  3. BE & HCO3-
  4. Simplified SID
  5. Lactate
  6. Anion gap
  7. Consider albumin

This entire article barely scrapes the surface for more in-depth acid-base analysis, and it doesn’t even cover differentials and causes for abnormalities. My main aim was to establish a grounding in the fundamentals. I’d suggest reading further from the references below to get a better understanding, if that’s the kind of thing you are after.

As always, all criticisms/corrections welcome.

References

  1. National Center for Biotechnology Information. (2019). “Stewart Acid-Base Balance.” In StatPearls. Available at: https://www.ncbi.nlm.nih.gov/books/NBK507807/
  2. Anaesthesia MCQ. “Acid-Base Physiology.” Available at: http://www.anaesthesiamcq.com/AcidBaseBook/ABindex.php
  3. Acid-Base. “The Acid-Base Physiology Community.” Available at: https://acidbase.org
  4. Weingart, S. “AG Metabolic Acidosis.” Internet Book of Critical Care (IBCC). EmCrit Project. Available at: https://emcrit.org/ibcc/agma/
  5. Weingart, S. (2011). “Understanding Acid-Base Disorders Using the Stewart Approach.” EmCrit Project. Available at: https://emcrit.org/wp-content/uploads/2011/04/acid-base-stewart.pdf
  6. Maseeh, A., & Saleh, A. (2014). “Facing acid-base disorders in the third millennium – the Stewart approach revisited.” International Journal of Nephrology and Renovascular Disease, 7, 349-354. doi:10.2147/IJNRD.S61538
  7. Jones, D. “Stewart Acid-Base: A Simplified Bedside Approach.” The Open Mind – Anaesthesia and Analgesia. Available at: https://www.openmindanaesthesia.org/stewart-acid-base-bedside-approach
  8. The Physiological Society. (2020). “An Introduction to Stewart Acid-Base.” The Physiological Society Magazine. Available at: https://www.physoc.org/magazine-articles/an-introduction-to-stewart-acid-base/
  9. National Center for Biotechnology Information. (2021). “Metabolic Acidosis in Critical Care and Anesthesia.” In StatPearls. Available at: https://www.ncbi.nlm.nih.gov/books/NBK559139/
  10. Park, M.A.J., Cave, G., & Freebairn, R.C. (2020). “Metabolic Acidosis in Anaesthesia and Critical Care.”Anaesthesia Intensive Care.
  11. Moviat, M.A.M., van Haren, F.M.P., & van der Hoeven, J.G. (2003). “Stewart’s approach: Just a heresy or another lens into acid-base physiology?” Critical Care, 7(3), 245-246.
  12. Morgan, T.J. (2009). “Acid-base analysis: a critique of the Stewart and bicarbonate-centered approaches.”Anaesthesia Intensive Care, 37(4), 482-487.
  13. Fencl, V., & Leith, D.E. (1993). “Bench-to-bedside review: Fundamental principles of acid-base physiology.” Critical Care, 7(3), 160-164.
  14. Deranged Physiology. “Traditional and Physico-chemical Approaches to Acid-Base.” Deranged Physiology – CICM Primary Exam Required Reading. Available at: https://derangedphysiology.com/main/cicm-primary-exam/required-reading/acid-base-physiology/Chapter%20412/traditional-and-physico-chemical-approaches-acid-base