How do I interpret arterial blood gas (ABG) results and calculate acid‑base disturbances, including compensation, anion gap, and mixed disorders?

Systematic ABG Interpretation Using a Stepwise Algorithm

Interpret arterial blood gas results using a systematic 5-step approach: (1) assess oxygenation and pH, (2) identify the primary disorder (respiratory vs metabolic), (3) determine if compensation is appropriate, (4) calculate anion gap for metabolic acidosis, and (5) evaluate for mixed disorders.

Step 1: Initial Assessment

Start by examining three critical values:

• pH: Normal 7.35-7.45
  • <7.35 = acidemia

  • 7.45 = alkalemia

• PaCO₂: Normal 35-45 mmHg
• HCO₃⁻: Normal 22-26 mEq/L

Also assess oxygenation adequacy 1:

• Target SpO₂ 94-98% for most patients
• Target SpO₂ 88-92% for those at risk of hypercapnic respiratory failure (COPD, severe obesity, neuromuscular disease)

Step 2: Identify Primary Disorder

Use the "RoMe" technique 2:

• Respiratory opposite: If pH and PaCO₂ move in opposite directions → respiratory disorder
  • pH ↓ + PaCO₂ ↑ = respiratory acidosis
  • pH ↑ + PaCO₂ ↓ = respiratory alkalosis
• Metabolic equal: If pH and HCO₃⁻ move in the same direction → metabolic disorder
  • pH ↓ + HCO₃⁻ ↓ = metabolic acidosis
  • pH ↑ + HCO₃⁻ ↑ = metabolic alkalosis

Step 3: Assess Compensation

For Metabolic Acidosis 3:

• Expected PaCO₂ = 1.5 × (HCO₃⁻) + 8 (±2)
• If actual PaCO₂ is higher than expected → concurrent respiratory acidosis
• If actual PaCO₂ is lower than expected → concurrent respiratory alkalosis

For Metabolic Alkalosis 3:

• Expected PaCO₂ increase = 0.7 × (HCO₃⁻ increase above 24)

For Respiratory Acidosis 3:

• Acute: HCO₃⁻ increases 1 mEq/L per 10 mmHg PaCO₂ rise
• Chronic: HCO₃⁻ increases 3.5 mEq/L per 10 mmHg PaCO₂ rise

For Respiratory Alkalosis 3:

• Acute: HCO₃⁻ decreases 2 mEq/L per 10 mmHg PaCO₂ fall
• Chronic: HCO₃⁻ decreases 5 mEq/L per 10 mmHg PaCO₂ fall

Deviations from expected compensation indicate mixed disorders 4, 5.

Step 4: Calculate Anion Gap (for Metabolic Acidosis)

Anion Gap = Na⁺ - (Cl⁻ + HCO₃⁻)

Normal range: 8-12 mEq/L (varies by institution) 6

Correct for albumin 7:

• For every 1 g/dL decrease in albumin below 4 g/dL, add 2.5 to the calculated anion gap

High Anion Gap Metabolic Acidosis (>12)

Use GOLDMARK mnemonic 3:

• Glycols (ethylene, propylene)
• Oxoproline
• L-lactate
• D-lactate
• Methanol
• Aspirin (salicylates)
• Renal failure (uremia)
• Ketoacidosis (diabetic, alcoholic, starvation)

Perform gap-gap analysis 3:

• Delta gap = (Calculated AG - Normal AG) - (24 - Measured HCO₃⁻)
• If delta gap >6 → concurrent metabolic alkalosis
• If delta gap <-6 → concurrent non-anion gap metabolic acidosis

Calculate osmolal gap if toxic ingestion suspected 3:

• Osmolal gap = Measured osmolality - Calculated osmolality
• Calculated osmolality = 2(Na⁺) + (Glucose/18) + (BUN/2.8)
• Gap >10 suggests toxic alcohol ingestion

Normal Anion Gap Metabolic Acidosis

Calculate urine anion gap 3:

• Urine AG = (Na⁺ + K⁺) - Cl⁻
• Positive (>0) → renal cause (RTA)
• Negative (<0) → GI bicarbonate loss (diarrhea)

Step 5: Evaluate for Mixed Disorders

Key indicators of mixed disorders 4, 5:

• pH near normal with abnormal PaCO₂ and HCO₃⁻
• Compensation exceeds expected values
• Abnormal delta gap
• Clinical picture doesn't match single disorder

Critical Clinical Thresholds

For patients with respiratory acidosis 8, 1:

• pH <7.35 with PaCO₂ ≥6.5 kPa (49 mmHg) and RR >23 after 1 hour of optimal therapy → initiate NIV (Grade A recommendation)
• For PaCO₂ 6.0-6.5 kPa (45-49 mmHg), consider NIV based on clinical context

Common Pitfalls

• Ignoring albumin correction leads to missing high anion gap acidosis in hypoalbuminemic patients 7
• Failing to calculate delta gap misses coexisting metabolic alkalosis in anion gap acidosis 3
• Not recognizing mixed disorders when compensation seems "too good" 4
• Overlooking the clinical context - ABG interpretation must align with patient presentation 3

The anion gap has proven value beyond acid-base diagnosis, serving as a significant predictor of morbidity and mortality across multiple clinical conditions 6.