Arterial blood gas test: Procedure, normal values, and interpretation

Key takeaways

An arterial blood gas test measures pH, PaO2, PaCO2, HCO3, and SaO2, plus base excess as a derived sixth value.

Normal pH runs 7.35 to 7.45, and anything outside that band needs urgent clinical assessment.

The radial artery is the first-choice site, and the modified Allen test comes before every radial puncture.

Read the result in five steps, starting with pH and ending with the clinical context.

Practice management software like Pabau keeps the order, the result, and the follow-up in one patient record.

A pulse oximeter tells you how saturated the blood is. An arterial blood gas test tells you why. It draws blood straight from an artery and reports six values at once, so you can see respiratory and metabolic function in the same snapshot.

That difference decides cases. A patient on supplemental oxygen can hold a reassuring SpO2 of 96% while carbon dioxide quietly climbs, and only an arterial sample will catch it. According to NCBI Bookshelf, blood gas analysis remains one of the most useful tests in emergency and critical care. It reads both systems from a single draw.

Getting value out of it depends on two things. You draw the sample cleanly, then read the numbers in a fixed order. Both are learnable, and both go wrong in predictable ways.

What the six arterial blood gas values tell you

Every arterial blood gas report returns the same core dataset. The analyzer measures pH, carbon dioxide, and oxygen tension directly. It then derives the rest, including base excess, which most reports print as a sixth value underneath.

Those ranges come from the American Thoracic Society and NCBI Bookshelf. Treat them as benchmarks rather than absolutes. Altitude and analyzer calibration shift the figures slightly, the same caveat that applies to any abnormal vital signs chart.

When an arterial blood gas test is worth the puncture

Order an arterial blood gas test when you need a measured number, not an estimate, for ventilation, oxygenation, or acid-base balance. This is not a routine screen, and it hurts more than a venous draw. So the indication has to justify the needle.

• Respiratory failure: acute or chronic hypoxemia, hypercapnia, or distress where the degree of impairment needs quantifying
• COPD exacerbation: to separate type 1 from type 2 failure and guide oxygen titration
• Metabolic emergencies: diabetic ketoacidosis, lactic acidosis, and renal failure, where acid-base status directs treatment
• Ventilator monitoring: serial gases guide pressure, rate, and oxygen settings
• Unexplained drowsiness or confusion: a gas can reveal CO2 retention or severe metabolic acidosis as the cause
• Perioperative assessment: thoracic surgery, or any patient with known pulmonary disease
• Toxicology: carbon monoxide poisoning, salicylate overdose, and mixed acid-base pictures

Two of those account for most gases drawn on a medical ward. A patient in respiratory failure needs serial samples to show whether oxygen therapy is working. A COPD exacerbation needs one early gas before oxygen is titrated, which is also why a COPD care plan opens with a documented baseline.

Meanwhile, MedlinePlus notes the test earns its place whenever pulse oximetry is unreliable. Carbon monoxide poisoning is the classic example, because SpO2 reads falsely normal while the patient deteriorates.

Outpatient settings order fewer gases, but they do order them. Infusion suites and IV therapy practices escalating a deteriorating patient may need one before transfer. The result then belongs in the same record as the infusion notes, not on a loose printout.

How to collect an ABG sample safely, step by step

Arterial sampling carries more risk than venous phlebotomy, so the sequence matters. Choose the site first, confirm the collateral circulation second, and only then puncture. Most clinicians work at the radial artery, with brachial and femoral sites held in reserve.

Radial first, brachial second, femoral last

The modified Allen test comes before every radial puncture

The modified Allen test confirms that the ulnar artery can supply the hand on its own. A negative result rules out radial sampling at that wrist, so pick another site instead of pressing on.

1. Compress the radial and ulnar arteries together while the patient opens and closes the fist several times
2. Ask the patient to open the hand, and watch the palm blanch
3. Release pressure over the ulnar artery only
4. Watch the color return, because flushing within 5 to 7 seconds means adequate ulnar flow
5. Treat flushing beyond 10 to 15 seconds, or none at all, as a negative test and choose another site

Once the test is positive, prep the site with antiseptic and load a pre-heparinized syringe. Enter the skin bevel-up at 30 to 45 degrees for a radial puncture. Brachial puncture is the site usually taught at a steeper 45 to 60 degrees.

After the draw, hold firm pressure for at least five minutes, and longer if the patient is anticoagulated. That single step prevents most hematomas. It is the same post-procedure discipline described in guidance on IV therapy complications, where recording what happened matters as much as doing it well.

A pre-puncture checklist worth running every time

Most avoidable problems start before the needle moves. Run through these eight points first, and the sample you send will be interpretable when it arrives.

• Write the indication in the record before you draw, not after the event
• Check anticoagulants, platelet count, and any bleeding history
• Run the modified Allen test and note the result
• Ask about a fistula, graft, or recent line in that arm, and switch sides if you find one
• Wait 20 to 30 minutes after any change to oxygen flow or ventilator settings
• Use a pre-heparinized syringe, and expel every air bubble before you cap it
• Label the sample with the time, the site, and the inspired oxygen concentration
• Get it to the analyzer within 15 minutes, or put it on ice

Read an arterial blood gas result in five steps

Work through the report in the same order every time, and mixed disorders stop hiding. The sequence runs from the broadest question to the most specific, which is what stops you fixating on one abnormal number.

1. Assess pH. Below 7.35 is acidemia, above 7.45 is alkalemia. This gives you the net state, not the cause.
2. Identify the primary disorder. With acidemia, a high PaCO2 means respiratory acidosis and a low HCO3 means metabolic acidosis. With alkalemia, reverse both.
3. Assess compensation. Has the other system responded? Compensation moves in the same direction as the primary change and never fully corrects the pH.
4. Assess oxygenation. Check PaO2 and SaO2, then grade any hypoxemia as mild at 60 to 79 mmHg, moderate at 40 to 59, or severe below 40.
5. Integrate the clinical context. A PaCO2 of 50 mmHg means one thing in a COPD patient on home oxygen and something else entirely after surgery.

A worked example from a night shift

A 68-year-old with known COPD arrives breathless on 4 L/min of oxygen. The gas reads pH 7.28, PaCO2 68 mmHg, HCO3 31 mEq/L, and PaO2 58 mmHg. Running the five steps takes about 20 seconds.

Start with the pH, which says acidemia. A raised PaCO2 makes that respiratory acidosis. Because the HCO3 is raised too, you know the retention is not new. That PaO2 sits in the moderate hypoxemia band. Put together, this is an acute-on-chronic exacerbation, documented and coded J44.1, with an oxygen saturation target of 88 to 92%.

Numbers serve the clinical story, never the reverse. That principle sits at the center of interpreting clinical biomarkers honestly, and it applies just as much at 3am. Nursing frameworks such as an impaired gas exchange plan exist to turn a result like that into an intervention and a review time.

Four acid-base patterns worth knowing on sight

Each primary disorder leaves a distinctive fingerprint on the report. Match the pattern first, then work out the cause, because the pattern narrows the differential faster than the history does.

Compensation is always partial in an acute presentation. So a pH that has returned fully to normal points to one of two things. Either the state is chronic and compensated, or two disorders are canceling each other out.

Choosing between an ABG, pulse oximetry, and a venous gas

Pick the test that answers the question in front of you. Pulse oximetry trends oxygenation continuously, a venous gas screens acid-base status, and only an arterial sample reports both properly.

If the choice still feels close at the bedside, three questions settle it.

• Is the oxygen therapy enough? Pulse oximetry answers that, unless perfusion is poor.
• Is the patient retaining carbon dioxide? Only an arterial sample answers that.
• Do you need a rough pH in a stable patient with difficult arterial access? A venous gas will do.

The same triage logic shows up with other time-critical blood markers. A troponin levels chart earns its place in a chest-pain workup for the same reason, because it converts a raw number into a decision.

Arterial puncture risks and how to keep them rare

Most complications of arterial puncture are minor and settle on their own. The ones worth worrying about are largely preventable with pressure, sensible site choice, and sterile technique.

• Hematoma: the most common problem, and the one firm pressure for five minutes prevents
• Arterial spasm: more common at the radial site, and usually transient
• Thrombosis or occlusion: rare, and likelier with peripheral vascular disease or repeated punctures at one site
• Infection: rare with sterile technique, and a higher risk in immunocompromised patients
• Nerve injury: greatest at the brachial site, where the median nerve runs beside the artery
• Vasovagal response: syncope or near-syncope, more common in anxious or pain-sensitive patients
• Air embolism: prevented by clearing every bubble from the syringe before sealing it

Radial puncture is off the table with a negative Allen test. The same goes for infection or burns over the site, or an arteriovenous fistula in that arm. Anticoagulated patients need 10 minutes of pressure or more. Document each of those decisions, because they are exactly what gets questioned later.

Where the sample runs changes what you can do with it

A bedside analyzer returns a result in two to five minutes, while a central lab takes 15 to 60. In a deteriorating patient, that difference decides whether the gas guides treatment or simply confirms what already happened.

Pre-analytical errors produce more wrong results than analyzer faults do. Four of them recur constantly, and every one is preventable.

• Air bubbles left in the syringe raise the PaO2 and lower the PaCO2
• Too much heparin dilutes the sample and drags the PaCO2 down
• A delay past 15 minutes at room temperature lets cells keep using oxygen, so the PaO2 falls
• A missing inspired-oxygen figure on the label leaves the PaO2 impossible to interpret

How Pabau keeps ABG orders and results in one record

In most practices, the gas result lands in one place and the reasoning behind it lands in another. The printout gets scanned, the indication sits in a note, and the follow-up lives in somebody’s head. That works until a different clinician reviews the patient, or a complaint arrives 18 months later.

Practices that order blood work regularly feel it first. Good GP practice software treats a result as part of the record, rather than an attachment sitting beside it.

Practice management software like Pabau files the result against the consultation that prompted it. The indication, the ordering clinician, the collection site and time, and the immediate response all sit in the same patient record.

That trail supports continuity of care, and it is the backbone of safer clinical notes when a case gets reviewed.

Documentation time is the other half of the problem. Pabau Scribe, our AI scribe, drafts the clinical note from the consultation itself. So the reasoning behind an urgent gas gets captured while it is fresh, rather than at the end of a long shift.

Teams comparing HIPAA compliant AI tools should check one thing above all. Does the tool write into the patient record, or into a document somebody then has to file?

Consent and pre-procedure checks belong in the same file as the result. Digital intake forms capture the Allen test outcome, the site chosen, and the risks you discussed. Joining the order to the outcome is where software stops being a filing cabinet rather than a record.

Keep test orders and results in one record

Pabau files every result against the consultation that prompted it, with the indication, the clinician, and the follow-up in the same patient record. Nothing sits in a separate folder waiting to be found.

Book a demo

Conclusion

An arterial blood gas test is only as good as the sample and the sequence. Draw it cleanly, run the five steps in order, and the result tells you what the lungs and kidneys are doing right now. Skip either discipline and you get a number that misleads with confidence.

The clinical side of this is taught well. The documentation side rarely is, and that is where practices lose both time and defensibility. Decide now where a gas result lives, who reviews it, and how the follow-up gets recorded.

Book a demo to see how Pabau keeps test orders, results, and clinical notes together in one patient record.

Continue your research

Need a care plan for a patient who cannot clear secretions? Ineffective airway clearance care plan sets out assessment findings, interventions, and review points you can document straight away.

Recording observations on paper at the bedside? Vital signs record gives you a structured sheet for tracking trends between blood gas samples.

Interpreting an unfamiliar blood panel? Rheumatoid arthritis test walks through what RF and anti-CCP results do and do not prove.

Explaining reference ranges to a patient? Blood sugar levels chart shows how normal, prediabetes, and diabetes bands are drawn and where they overlap.

Coding an abnormal respiratory specimen result? ICD-10 code R84.4 covers documentation and billing for abnormal immunological findings.

Frequently asked questions