Do we perform an arterial blood gas (ABG) if oxygen saturation (SpO2) is low?

When to Perform ABG with Low SpO2

Yes, perform an arterial blood gas (ABG) when SpO2 is low, specifically when SpO2 falls below 94% in most patients or below 88-92% in those at risk for hypercapnic respiratory failure, particularly if there is clinical deterioration, unexplained symptoms, or need for increased oxygen therapy. 1

Clinical Decision Algorithm for ABG Testing

Immediate ABG Indications (Perform Within 1 Hour)

• SpO2 <85%: Obtain ABG immediately after initiating oxygen therapy to assess for hypercapnia and guide management 2
• Critical illness or shock (systolic BP <90 mmHg): Use arterial sampling rather than capillary samples 1
• Unexpected SpO2 fall ≥3% below baseline or any inappropriate fall below 94% in patients breathing room air or supplemental oxygen 1

Standard ABG Indications with Low SpO2

• SpO2 <94% in patients without risk factors for CO2 retention: Perform ABG to confirm hypoxemia and assess acid-base status 1, 3
• SpO2 <96% in patients with suspected hepatopulmonary syndrome: ABG is required for diagnosis, as SpO2 <96% has 100% sensitivity and 88% specificity for detecting HPS 1
• Any patient requiring increased FiO2 to maintain constant oxygen saturation: ABG needed to assess for worsening gas exchange 1

Special Populations Requiring ABG

• Patients at risk for hypercapnic respiratory failure (COPD, severe obesity, neuromuscular disease, chest wall deformities): Obtain ABG when SpO2 drops below their target range of 88-92% 1, 4
• Patients with unexplained confusion, agitation, or drowsiness: These may indicate hypercapnia even with adequate SpO2 readings 1, 3
• Deteriorating patients with metabolic risk factors: Diabetic ketoacidosis or renal failure with metabolic acidosis 1

Critical Limitations of Pulse Oximetry

Why SpO2 Alone Is Insufficient

• Normal SpO2 does not exclude serious pathology: A patient can have normal SpO2 but abnormal pH, PCO2, or low oxygen content from anemia 1
• Pulse oximetry overestimates arterial saturation: Mean difference of 2.75% in septic patients, with 50% of patients showing SpO2 90-93% actually having SaO2 <90% 5
• Accuracy decreases with hypoxemia: When SaO2 <90%, pulse oximetry overestimates by mean of 4.9% compared to 1.89% in non-hypoxemic patients 5
• Poor reliability in critically ill patients: Standard deviation of differences between SpO2 and SaO2 ranges from 2-3%, with limits of agreement up to ±6% 6, 7

Factors That Worsen Pulse Oximetry Accuracy

• Hypoxemia (SaO2 <90%): Significantly reduces accuracy 5, 8
• Vasoactive drug use: Decreases precision of measurements 6, 5
• Poor perfusion states: Low perfusion index provides only marginal improvement in accuracy 7
• Acidosis and hyperlactatemia: Worsen precision though not necessarily bias 5

Clinical Pitfalls to Avoid

• Do not rely on SpO2 alone in critically ill patients: ABG is essential before adjusting FiO2 settings in ICU patients 7
• High SpO2 thresholds needed for safety: SpO2 >94% is necessary to ensure SaO2 ≥90% with good sensitivity 6
• Beware of false reassurance: In hepatopulmonary syndrome, SpO2 96% may seem adequate but requires ABG for proper diagnosis 1
• Consider clinical context: Tachypnea (>30 breaths/min) or increased work of breathing requires ABG even with adequate SpO2 3, 4

When ABG May Not Be Immediately Necessary

• Stable patients with SpO2 ≥94% without risk factors for CO2 retention and no clinical deterioration: Simple monitoring is appropriate 3
• Transient dips to 90% during sleep in otherwise normal individuals: May not require ABG 1
• After initial ABG confirms no hypercapnia: Subsequent monitoring may rely more on pulse oximetry with clinical correlation 1