ABG Interpretation Trainer
Compensation essentially never normalises the pH — so a normal pH with abnormal chemistry is a mixed disorder, not a compensated one.
A study aid, not a diagnostic tool. Compensation rules are population regressions with real scatter, derived from patients with single established disorders in a steady state that takes hours to days to reach. A gas is a snapshot of a moving process, and no algorithm interprets one — the history does most of the work. Nothing here knows anything about a patient.
What full compensation actually achieves
A gas is two measurements and a calculation
Henderson-Hasselbalch ties the three values together — pH equals 6.1 plus the log of bicarbonate over 0.03 times pCO₂ — so given any two, the third is determined. Most analysers measure pH and pCO₂ and compute the bicarbonate, which means a reported set that does not satisfy the equation is an error somewhere rather than an interesting physiological finding. Checking that consistency takes seconds and catches transcription slips, venous samples labelled arterial, and delays before analysis.
The single most useful fact in acid-base is that compensation essentially never normalises the pH. Sweeping metabolic acidosis with Winter's formula, one of nineteen bicarbonate values produces a normal pH — and that one is a bicarbonate of 22, which sits at the bottom of the normal range and is barely deranged to begin with. Metabolic alkalosis with its expected compensation normalises none of thirteen. So a normal pH alongside a clearly abnormal pCO₂ and bicarbonate is a mixed disorder, not a triumphantly compensated simple one. The body drags the pH partway and stops.
There is one documented exception, and it is worth knowing precisely because it is the case where a normal pH does not imply a second disorder: chronic respiratory alkalosis genuinely can return the pH to normal — at a pCO₂ of 30 with a fully chronic bicarbonate of 20 the pH is 7.45. That is why the rule is stated as "essentially never" rather than "never". And the acute-versus-chronic question cannot be answered from the gas at all: at a pCO₂ of 60 the two expectations differ by 5 mmol/L of bicarbonate, so the same numbers read either way depending entirely on a history the analyser never sees.
How to use
- Enter the pH, pCO2 and bicarbonate from the gas.
- Check internal consistency first — it takes seconds and catches errors.
- Compare the measured compensation with the expected value.
- Decide acute or chronic from the history, not from the gas.
Frequently asked questions
Are the three values on a blood gas independent?
No. Henderson-Hasselbalch ties them together — pH equals 6.1 plus the log of bicarbonate over 0.03 times pCO2 — so given any two, the third is determined. Most analysers measure pH and pCO2 and compute the bicarbonate, which means a gas is really two measurements and a calculation rather than three measurements.
How can I tell if a blood gas result is wrong?
Check that the three values satisfy Henderson-Hasselbalch. If the reported pH differs from the one the equation implies by more than about 0.05, something is wrong — a transcription slip, a venous sample labelled arterial, or a delay before analysis. The equation admits no exceptions, so an inconsistency is an error rather than a finding.
Does compensation ever return the pH to normal?
Essentially never, and this is the single most useful fact in acid-base. Sweeping metabolic acidosis with Winter’s formula, one of nineteen bicarbonate values gives a normal pH — and that one is a bicarbonate of 22, at the bottom of the normal range and barely deranged. Metabolic alkalosis with expected compensation normalises none of thirteen.
What does a normal pH with an abnormal pCO2 and bicarbonate mean?
A mixed disorder, in almost every case. Since compensation drags the pH partway and stops rather than overshooting back to normal, a normal pH alongside two clearly deranged values points at two processes pulling in opposite directions. Reading it as a triumphantly compensated simple disorder is the classic error.
Is there an exception to compensation not normalising the pH?
One, and it is documented: chronic respiratory alkalosis can genuinely return the pH to normal. At a pCO2 of 30 with a fully chronic bicarbonate of 20 the pH is 7.45, inside the range. It matters precisely because it is the single case where a normal pH does not imply a second disorder, and it is why the rule is stated as essentially never rather than never.
What is Winter’s formula?
The expected pCO2 in a metabolic acidosis: 1.5 times the bicarbonate plus 8, within about 2 either way. A measured pCO2 above that suggests an additional respiratory acidosis, and below it an additional respiratory alkalosis. Compensation that looks excessive is usually a second disorder rather than unusual enthusiasm.
How do I tell acute from chronic respiratory failure on a gas?
You cannot, and that is the honest answer. The expected bicarbonate differs substantially — at a pCO2 of 60 an acute picture expects about 26 and a chronic one about 31, five apart — so the same gas reads either way depending on which you assume. Only the history settles it, and a chronic retainer presenting acutely unwell has both pictures at once.
How much does the bicarbonate rise in respiratory acidosis?
About 1 mmol/L per 10 mmHg of pCO2 rise acutely, and about 3.5 per 10 once chronic, because the renal response takes days to develop. In respiratory alkalosis it falls about 2 per 10 acutely and 4 per 10 chronically. Those are population figures with real scatter rather than precise laws.
Why does compensation take time?
Because the two systems work at very different speeds. The respiratory response to a metabolic problem begins within minutes and is largely complete in hours; the renal response to a respiratory problem takes two to five days. That is the entire basis of the acute-versus-chronic distinction, and it is why a gas taken mid-change fits none of the rules.
What does appropriate compensation tell me?
That the compensating system is intact — not that the problem is resolved or that the patient is well. A perfectly compensated metabolic acidosis still has a low pH and whatever caused it is still there. Compensation is evidence about the physiology rather than reassurance about the patient.
Why do compensation rules fail at the edges?
Because they are regressions fitted to patients with single, established disorders in a steady state, and they were never intended as hard boundaries. Treating a rule tolerance as a threshold produces confident nonsense in exactly the complicated patients where the answer matters, which is where the history has to take over.
Can I interpret a real gas with this?
No. This is a study and teaching aid for people learning the arithmetic and the traps in it. Reference ranges differ between laboratories, a gas is a snapshot of a moving process, and the same numbers mean different things in chronic lung disease, in vomiting and in a patient who has just been sedated. No algorithm interprets a gas.
🔒 This tool runs entirely in your browser. Nothing you enter is uploaded, logged, or stored.