ABG Analysis


NOTE: Before reading this, you need to be confident with reading ABGs. If you’re not, click here for a whole website you can practice on.

In this page, we’ll look at:

What is blood gas analysis? ¶

When you look at an ABG and call it “respiratory acidosis partially compensated by a metabolic alkalosis,” you are describing the results. But to react clinically, you need to deduce how the patient got to that state. That process is ABG analysis.

Remember that in an ABG, the pH tells you where your patient is; the CO2 and bicarb tell you how the patient got there. This is important. What do you do if your patient has this blood gas?


The correct answer is: ask what they’d like for lunch. The pH is where your patient is, and this patient is at bang-on normal. The only way to “fix” it is to take away their perfect pH. This patient lives at those levels. You can work on lifestyle choices, but don’t try to make any acute changes!

A hands-on experiment ¶

Everything about analysis is based on two simple facts:

The respiratory component (CO2 ) changes quickly, breath-by-breath.

The metabolic component (bicarb) changes slowly over the course of a couple days.

Try a little experiment. Look at your watch and pant as hard as you can for 10 seconds. Seriously, do it.


Do it.

By second 4 you were probably getting anxious, and by second 10 you probably felt like crap. You caused yourself to go into a respiratory alkalosis (CO2 changes quickly!), and it’s uncompensated because metabolic compensation takes a couple days.

Now imagine panting like that for a couple days, non-stop, to give your kidneys enough time to compensate with bicarb. Is it starting to make sense why a respiratory alkalosis is usually not the cause of a compensated blood gas?

Here’s the point of all this. Take a look at this blood gas:


There are two ways of reading this. This is either a fully compensated respiratory alkalosis, or a fully compensated metabolic acidosis. Which do you think is more likely?

Acute changes ¶

Acute changes present differently in a blood gas depending on whether the aetiology is metabolic or respiratory. Why? Because the respiratory system will compensate quickly, but the metabolic system won’t.

What might have caused this?


We have an uncompensated respiratory acidosis here, so this patient has started hypoventilating. But the real question is: how long has this been going on for? Is this more likely a COPDer or a narcotic intoxication?

The lack of metabolic compensation should suggest the timeframe. If this were chronic and there had been enough time for compensation, what would you expect the pH and bicarb to be?

Now, think about a patient having an asthma attack. They feel their airways beginning to tighten, and that is incredibly scary. The first thing they do is panic, so they become temporarily tachypneic and hyperventilate. Their initial ABG might look like this:


Think back to the blood gas we looked at after you hyperventilated. We said that a compensated respiratory alkalosis was unlikely to occur, but uncompensated is exactly what we’re looking at here. Uncompensated respiratory alkalosis could be an early warning sign of anxiety, pain, fear, a head injury (as in central neurogenic hyperventilation), or metabolic derangement.

After a bit of time, bronchoconstriction sets in and they begin gas trapping (that is, hypoventilating and retaining CO2 ). At first, their blood gas may look like this (why?):


Do you see why a “normal” blood gas can be completely misleading? The blood gas is normal (going between different pathologic states), but the patient isn’t. Once they start getting tired, the gas might look like this:


Take a close look at what we have here. The pH is acidotic, the CO2 is acidotic, and the bicarb is WNL. In other words, a change occurred, and we suspect it’s new because there is no compensation present. That leads us to this definition:

An acute respiratory blood gas shows a respiratory change (CO2 and pH) without metabolic compensation.

Metabolic changes are more complex, because the respiratory system can compensate quickly.

Diabetic ketoacidosis is the classic example of a metabolic acidosis. Fixed acids (those that can’t be exhaled) build up in the bloodstream. What breathing pattern is commonly associated with DKA? What do Kussmaul respirations look like, and how will that change the ABG? Remember: more breathing = more CO2 gets vented.

As another example, think about renal failure. Bicarb is modulated by the kidneys (as well as in the blood, but ignore that for now). If the kidneys stop kidneying, they can’t excrete acids, or reclaim or synthesize bicarbonate, so HCO3- levels drop.

Or think about shock states like sepsis. Anaerobic metabolism produces lactic acid as a byproduct, which builds up in the bloodstream.

All these things cause serum bicarb levels to drop.

It’s important to remember that the brain doesn’t actually respond to CO2 , only to H+ ions. We pretend that “CO2 makes you breathe” as a gross oversimplification, because CO2 is converted to H+ in the CSF. There’s no difference to your brain between high CO2 and a buildup of fixed acids or a drop in bicarb.

So. You probably won’t ever see this blood gas. Why?


H+ ions are building up, because there isn’t enough bicarb around to buffer them. Uncompensated metabolic acidosis is not what you’d expect in a patient, because the brain senses the acid load and immediately begins to hyperventilate. So, you’re more likely to see this:


When you look at this gas, why should you think partially compensated metabolic acidosis, rather than partially compensated respiratory alkalosis?

An acute metabolic blood gas shows change in both pH and bicarb, often with some degree of respiratory compensation.

If I saw the prior blood gas, with the normal CO2 (40), I’d be very worried that my patient was tiring out and I’d start considering BiPAP or intubation.

Think about that last statement.

If the respiratory system isn’t compensating for an acute metabolic change, your patient may be headed for a crash.

Combined disorders (respiratory and metabolic) are terrifying to people learning blood gasses. They make initial interpretation difficult (our brain looks for compensation, because that’s what we’re used to seeing), but they are perfectly common states.

Death (cardiac arrest) is a great example of a mixed condition.


Tissue hypoxia leads to massive lactic (metabolic) acidosis, and dead folks don’t breathe good (respiratory acidosis).

Chronic changes ¶

Patients with COPD retain CO2 for a living. Their kidneys have compensated for the extra CO2 by increasing HCO3- levels, bringing the pH back to normal.


There are abnormal values there, but remember that pH is where your patient is. This is probably their normal. Don’t make acute changes!

We’re arriving at the definition of a chronic gas. “Abnormal values with a normal pH” is almost the definition, but with one very important piece missing.

A chronic blood gas shows fully-compensated abnormalities in a patient at their baseline.

That last bit is the key.

How do you differentiate between a (chronic) fully-compensated respiratory acidosis (COPD), and an (acute) fully-compensated metabolic alkalosis (say, food poisoning with a big acid loss from protracted vomiting)? By looking at your patient! Or in this case, smelling them, I guess.

Never ever reach a conclusion from an ABG without assessing the patient!

Let’s look at an ABG we had earlier:


In a renal patient, this could be acute kidney injury (AKI) or chronic kidney disease (CKD). How do you know whether this is their baseline or an acute change? By looking at your patient.

There’s another important point in this blood gas: yes, 7.34 is outside the normal range of pH. Let it go. Remember, CO2 changes breath-by-breath, so the pH will change breath-by-breath too. Plus, getting stabbed in the wrist for an ABG sucks. Unless you’re an ABG bad-ass, the mere act of obtaining the blood gas will alter the result. Eat your heart out, Heisenberg.

My ABG Ninja! interactive learning tool makes a completely erroneous assumption about compensation, because schools teach it and I’m obligated to utilize it or I’d screw students. The faulty assumption is that the body never overcompensates.

It only takes a few breaths to overcompensate past 7.40.

The following blood gas would be interpreted as a fully compensated respiratory alkalosis, based on the assumption that the body would never compensate past 7.40:


This is far more likely to represent a fully compensated metabolic acidosis where the patient took a couple large breaths when the gas was being drawn.

Remember, the pH is where the patient is. This patient is fully compensated, and which side of 7.40 they’re sitting at is pretty much irrelevant unless it’s at one of the extremes.

Acute-on-chronic changes ¶

Let’s go back to our COPDer. This was the blood gas that you took when they were at their baseline.


What happens when they have an acute exacerbation, usually from getting a pneumonia? Their CO2 will go up and their pH will go down, without a concomitant change in HCO3- .


Here, the CO2 rose and the pH fell, and the bicarb remains elevated because there hasn’t been enough time for it to compensate.

Remember, the pH is where your patient is. If their pH is outside the normal range, something acute is going on.

An acute-on-chronic blood gas shows an acute pattern (change in both pH and CO2 or bicarb), superimposed on a bicarb that isn’t fully explained by the acute process.

That was a mouthful.

Think about it this way: an abnormal pH means an acute process. If that acute process doesn’t fully explain the bicarb value you’re seeing, suspect an acute process on top of a chronic one.

If someone’s breathing has gone so wrong that their CO2 climbed to 71, they’re not going to wait 3 days before going to the hospital. That bicarb of 34 was likely there to begin with.

Advanced stuff ¶

Casual ABG interpreters look at ABGs and describe them (“partially compensated metabolic acidosis”), and that’s where the vast majority of health care workers operate. The most crucial skill for front-line staff is identifying when something is abnormal, so that level of ABG comprehension is right where they should be.

ABG analysis really is an advanced level of comprehension. If you understood the stuff on this page, then you are doing really well!

But when you study ABG analysis in depth, you learn that everything you ever knew about blood gasses is wrong. Don’t worry about any of this stuff if you’re working to get your head around the other stuff.

Here’s just a taste of why ABG interpretation is often incorrect.

Think back to the DKA from earlier. The ketoacids are “fixed” acids, meaning that they can’t be breathed off (only carbonic acid, \( \ce{H2CO3} \), can do that). They caused a metabolic acidosis, so the bicarb value went down. And it’s not just acids—if you drank a strong base like oven cleaner, your bicarb would go up (though it would go down shortly thereafter, once you died).

So if acids and bases both modify the bicarb levels, then what is the ABG analyzer actually measuring?

Nothing at all.

That’s right: analyzers don’t measure the metabolic system at all. They measure pH and PaCO2 , and calculate the bicarb according to a derivation of the Henderson-Hasselbach equation:

$$ \ce{HCO3-} = 0.03 \times \ce{CO2} \times 10^{\text{pH} - 6.1} $$

The reason that the “metabolic system” umbrella covers everything is that the only thing that’s actually getting measured is CO2 . All the other acids and bases still affect pH, but we don’t measure them and so we call them “metabolic.”

The upshot of this is that for any two values (CO2 and/or HCO3- and/or pH), the third value is completely deterministic. Experienced blood gassers can often hear two values and know where the third will be.

So what test actually measures bicarbonate? Ironically enough, Total CO2 , aka TCO2 . The majority of CO2 (about 95%) is carried in the blood as bicarbonate ions, so the total bicarbonate count is pretty close to the total CO2 count.

So how does CO2 become HCO3- ion? This equation is called the hydrolysis reaction, and it is the actual chemical relationship between CO2 and HCO3- ion:

$$ \ce{CO2 + H2O <=> H2CO3 <=> H+ + HCO3-} $$

When CO2 goes up, it drives the reaction to the right, which produces more bicarb (HCO3- ). In other words, when CO2 goes up, bicarb goes up too because chemistry, and that does NOT represent compensation! An acute CO2 increase of 10 automatically causes an HCO3- increase of 1. Consider an otherwise healthy patient whose ABG is now:


This isn’t actually an uncompensated respiratory acidosis. Assuming their CO2 went from a normal 40 to 50, their bicarb went from 21 to 22. This person is actually having a combined respiratory and metabolic acidosis!

An acute CO2 drop of 10 causes a bicarb drop of 2 (not 1!). And chronic CO2 changes lead to a bicarb increase of 4 or a drop of 5.

The hydrolysis reaction works in both directions, so changes in bicarb will change CO2 levels without representing compensation. A bicarb increase of 1 raises CO2 by 0.7, but a drop of 1 lowers CO2 by 1.2.

Why the asymmetry? Because exhaled CO2 is lost from the system. The majority of CO2 transported in the blood is stored as bicarbonate ions. When CO2 is added, some of it is converted to bicarb. When CO2 is exhaled, a greater amount of bicarb must be hydrolyzed to reach equilibrium. What does this have to do with understanding ABG analysis? Absolutely nothing.

In a pure metabolic acidosis, Winter’s formula gives the expected CO2 if the body were fully compensating:

$$ \ce{CO2} = ( \ce{HCO3-} \times 1.5 ) + 8 \pm 2 $$

Anything outside of that range could be considered a concomitant respiratory acidosis or alkalosis. So if your patient’s bicarb is 18, their ideal CO2 should be 33—37. In other words, if that patient had a perfect normal CO2 of 40, they’d actually have a perfectly abnormal respiratory acidosis. Ouch.

Next steps ¶

There are literally dozens of rules-of-thumb for ABG analysis, enough to make you question your own existence.

The best place to start is Dana Oakes’s ABG Pocket Guide. It has a solid foundation of ABG analysis, in a just-the-facts-ma’am presentation. If you really want to bring the pain, William J. Malley’s Clinical Blood Gases: Assessment and Intervention contains a wealth of information in a practical, accessible delivery.

I have some other ABG learning tools you may find useful.

If you found this page helpful, consider sending me an email, for no other reason than they make me feel good.