Arterial Blood Gas Interpretation: the method, the errors, and the exam

Why this skill decides answers

An ABG question rarely asks you to state a diagnosis outright. It gives four numbers and expects you to name the disturbance, judge the compensation, and pick an intervention that matches. Get the order of reading wrong and every downstream judgement is wrong too, even if you can recite normal ranges perfectly.

The reason a fixed sequence matters is that pH, CO2, and bicarbonate do not carry equal weight. pH tells you the direction of the problem. CO2 and bicarbonate tell you which system caused it. Read them out of order and you can talk yourself into the wrong system, especially on a mixed picture where both CO2 and bicarbonate are abnormal in the same direction.

This is why ABG interpretation sits at the centre of respiratory and metabolic content rather than as a standalone topic. Questions on COPD, DKA, sepsis, and post-operative ventilation all route through the same four-step read. Master the sequence once and it transfers everywhere a gas is drawn.

How to do it reliably

Step one: look at pH alone. Below 7.35 is acidosis, above 7.45 is alkalosis. Do not look at CO2 or bicarbonate yet. This single number tells you which direction the body has tipped, and every later step is judged against it.

Step two: look at CO2. Normal is 35 to 45 mmHg. If CO2 moves in the opposite direction to pH, that is, CO2 is high when pH is low, or CO2 is low when pH is high, the respiratory system is the cause. Step three: look at bicarbonate. Normal is 22 to 26 mEq/L. If bicarbonate moves in the same direction as pH, that is, bicarbonate is low when pH is low, or bicarbonate is high when pH is high, the metabolic system is the cause.

Step four: compensation. Check whether the system not responsible for the primary problem has started to move in a direction that would correct the pH. If pH is still abnormal, compensation is partial. If pH has returned to normal range while CO2 and bicarbonate are both still abnormal, compensation is full. This order, pH, then CO2, then bicarbonate, then compensation, works on every gas you will be given, because it never asks you to judge two variables at once.

The common errors

The most frequent error is reading CO2 before pH, usually because CO2 sits second on the printed lab slip. This flips the logic: instead of asking which system explains the pH, the reader starts guessing at a diagnosis before knowing the direction of the problem, and the guess is often wrong on mixed disorders.

A second error is confusing compensation with correction. Compensation is the body's attempt to normalise pH by moving the other system; it does not mean the pH is actually normal. A test item describing a patient with a low pH, high CO2, and high bicarbonate is showing partial compensation, not resolution, and still needs the primary problem treated.

A third error is treating bicarbonate as if it moves as fast as CO2. Respiratory compensation for a metabolic problem can appear within minutes, because the lungs adjust rate and depth immediately. Renal compensation for a respiratory problem takes one to three days, because the kidneys must alter bicarbonate reabsorption. A gas drawn twenty minutes after a metabolic acidosis will already show a compensatory drop in CO2; a gas drawn twenty minutes after acute respiratory failure will not yet show a bicarbonate change.

Drills that build it

Write the four-step order on a card and force yourself to cover the CO2 and bicarbonate values while you look at pH alone. Say the direction out loud before uncovering the rest. This breaks the habit of scanning all four numbers at once, which is where the ordering error creeps in.

Practise with gases that have no compensation at all, since these are the easiest to get wrong through overthinking. A pH of 7.28, CO2 of 55, and bicarbonate of 24 is uncompensated respiratory acidosis: the bicarbonate has not moved, so there is nothing to weigh against the CO2. Students who are used to hunting for compensation sometimes invent it where none exists.

Then move to fully compensated gases, where pH sits at the edge of normal, for example 7.36, with both CO2 and bicarbonate clearly abnormal. These are the ones that separate a nurse who has memorised the four steps from one who applies them, because the temptation is to call the gas normal and stop reading.

Exam application

NCLEX-style items give you the gas and the clinical picture together, and expect the interpretation to inform the intervention, not just the label. A COPD patient with a chronically compensated respiratory acidosis needs a different oxygen target than a patient in acute respiratory failure, even though both may show a low pH and high CO2 on a single reading taken in isolation.

Watch for items that give you a gas and a potassium level together. Acidosis shifts potassium out of cells and can produce a falsely elevated serum potassium; correcting the acidosis without accounting for this can drop potassium sharply. The question is testing whether you connect the ABG read to a second lab value, not just the four-step method in isolation.

Distractor answers on these items are usually built around getting step one wrong. An option describing bicarbonate administration for a patient in respiratory acidosis with adequate compensation is designed to catch someone who identified the CO2 problem but skipped judging pH first to see whether the patient actually needed intervention beyond supporting ventilation.

Quick reference

Normal ranges: pH 7.35 to 7.45, CO2 35 to 45 mmHg, bicarbonate 22 to 26 mEq/L. Read in that order, every time: pH for direction, CO2 for respiratory involvement, bicarbonate for metabolic involvement, then compensation for how far the body has corrected.

Respiratory acidosis: low pH, high CO2. Respiratory alkalosis: high pH, low CO2. Metabolic acidosis: low pH, low bicarbonate. Metabolic alkalosis: high pH, high bicarbonate. If both CO2 and bicarbonate are abnormal in the same direction as each other rather than opposite to pH, consider a mixed disorder rather than forcing a single-system answer.