11.7 Arterial Blood Gases and Acid-Base Interpretation

Why Acid-Base Is a Cardiac Topic

Test-plan item IV.B.1 is arterial blood gas analysis. On a cardiac unit the gas is rarely ordered to diagnose a lung disease. It is ordered to answer three cardiac questions: is perfusion adequate, is the ventilatory pump failing, and is the metabolic milieu one in which vasoactive drugs and the myocardium can work. A pH of 7.15 explains a norepinephrine infusion that has stopped working; a bicarbonate of 38 explains a torsades episode in a patient on high-dose furosemide.

Normal Values

A Four-Step Method You Can Actually Use

Step 1 — Read the pH. Below 7.35 is acidemia, above 7.45 is alkalemia. A pH inside the range does not mean normal: it can be fully compensated, or a mixed disorder whose components cancel.

Step 2 — Name the primary disorder. In a respiratory disorder the PaCO2 moves in the direction opposite to the pH. In a metabolic disorder the bicarbonate moves in the same direction as the pH. Whichever value explains the pH is the primary problem.

Step 3 — Test compensation with a formula. Compensation is predictable, which is what makes mixed disorders detectable. The two rules that catch the most errors: compensation is never complete for a metabolic disorder in the acute setting, and compensation never carries the pH past 7.40 onto the opposite side. If it appears to have done so, there are two primary disorders.

Step 4 — Assess oxygenation. PaO2, the P/F ratio (PaO2 divided by FiO2 as a decimal; 300 or less defines ARDS oxygenation with lower values grading severity), and the alveolar-arterial gradient, which distinguishes hypoventilation (normal gradient) from shunt and V/Q mismatch (widened gradient).

Expected Compensation Formulas

Worked Examples

1. Acute respiratory acidosis. pH 7.24, PaCO2 62, HCO3 26, PaO2 58. Post-procedural oversedation after cardioversion. The PaCO2 moved opposite the pH, so the disorder is respiratory. Expected acute rise in bicarbonate is 1 per 10, and the PaCO2 is 22 above 40, so 24 plus 2.2 equals about 26.2 — the measured 26 matches — matching, therefore acute and uncompensated. Nursing action: stimulate, support ventilation, consider reversal of the sedative, prepare for non-invasive ventilation.

2. Acute respiratory alkalosis. pH 7.52, PaCO2 27, HCO3 22. A patient two days after knee arthroplasty with sudden dyspnea and tachycardia. Expected acute fall in bicarbonate is 2 per 10, and the PaCO2 is 13 below 40, so 24 minus 2.6 equals about 21.4; the measured 22 sits right on it, confirming an acute process. In this clinical setting the gas is pointing at pulmonary embolism, not at anxiety.

3. Metabolic acidosis with appropriate compensation. pH 7.21, PaCO2 24, HCO3 10, sodium 138, chloride 100. Anion gap = 138 minus (100 + 10) = 28, a wide gap. Winter's: (1.5 x 10) + 8 = 23, plus or minus 2, so 21-25. The measured 24 fits, so respiratory compensation is appropriate and there is a single primary disorder. Clinical picture: cardiogenic shock with lactic acidosis.

4. Metabolic alkalosis. pH 7.52, PaCO2 48, HCO3 38, potassium 3.0, chloride 88. Four days of intravenous furosemide plus nasogastric suction. Expected PaCO2 = (0.7 x 38) + 20 = 46.6, plus or minus 5. The measured 48 is appropriate, so this is a compensated primary metabolic alkalosis.

5. Mixed disorder. pH 7.10, PaCO2 58, HCO3 17, during and after cardiac arrest. Both values move in the acidemic direction — the PaCO2 is high (acid) and the bicarbonate is low (acid). Values that move in the same acid-base direction can never be compensating for one another, so this is a combined respiratory and metabolic acidosis: inadequate ventilation and pulmonary blood flow layered on lactic acidosis. This pattern is the classic arrest gas, and its treatment is better perfusion and better ventilation, not bicarbonate.

6. The hidden-acidosis trap. pH 7.44, PaCO2 26, HCO3 17, sodium 140, chloride 100, in a septic patient who has also been vomiting. The near-normal pH with a low PaCO2 looks like a compensated respiratory alkalosis, but run the numbers: the PaCO2 is 14 below 40, so an acute respiratory alkalosis predicts a bicarbonate of about 24 minus 2.8, or 21.2. The measured 17 is well below that, so a metabolic acidosis is hiding underneath — and the anion gap of 140 minus (100 + 17) = 23 confirms it. Calculate the anion gap on every abnormal gas, because a normal pH never rules out a mixed disorder.

The Anion Gap

Anion gap = sodium minus (chloride + bicarbonate). The normal range is 8-12 mEq/L with older analysers and 3-11 mEq/L with many contemporary ion-selective electrodes; use your laboratory's stated normal.

Albumin correction is mandatory in critical care. Albumin is the dominant unmeasured anion, so hypoalbuminemia lowers the apparent gap. Corrected anion gap = measured gap + 2.5 x (4.0 minus the albumin in g/dL). A cardiac ICU patient with albumin 2.0 and an apparent gap of 12 actually has a corrected gap of about 17 — a lactic acidosis that would otherwise be missed entirely.

The Delta Ratio

Delta ratio = (measured anion gap minus normal anion gap) divided by (normal bicarbonate minus measured bicarbonate). It answers the question of whether the bicarbonate fall is fully explained by the gap.

High Anion Gap Acidosis in the Cardiac Patient

• Lactic acidosis. Type A from hypoperfusion: cardiogenic shock, cardiac arrest, hemorrhage, mesenteric ischemia, and limb ischemia from an intra-aortic balloon pump sheath or an ECMO arterial cannula. Type B without hypoperfusion: metformin (particularly once an acute kidney injury develops after contrast or aggressive diuresis), epinephrine and high-dose beta-2 agonists driving glycolysis, liver failure, thiamine deficiency, propofol infusion syndrome and linezolid. The scenario in which lactate rose after epinephrine was started is not necessarily worsening shock.
• Ketoacidosis, including euglycemic diabetic ketoacidosis on an SGLT2 inhibitor. Empagliflozin and dapagliflozin are now standard therapy across the ejection fraction spectrum, so this is a cardiac-unit problem. The patient has a wide-gap acidosis with positive beta-hydroxybutyrate and a glucose that may be only 150-200 mg/dL, which is exactly why it is missed. Precipitants are fasting, vomiting, acute illness and surgery. SGLT2 inhibitors are held 3 to 4 days before major elective surgery for this reason.
• Uremia from cardiorenal syndrome and advanced chronic kidney disease.
• Toxic alcohols (methanol, ethylene glycol), where an osmolar gap accompanies the anion gap, and salicylate, which classically produces a mixed high-gap acidosis with a primary respiratory alkalosis.

Non-Gap (Hyperchloremic) Acidosis

• Diarrhea or high-output ileostomy losing bicarbonate
• Type 4 renal tubular acidosis with hyperkalemia — extremely common in cardiac patients on ACE inhibitors, angiotensin receptor blockers or mineralocorticoid receptor antagonists, especially with diabetes
• Acetazolamide
• Large-volume 0.9% sodium chloride. The supraphysiologic chloride load displaces bicarbonate. Balanced crystalloids such as lactated Ringer's or Plasma-Lyte avoid this and are preferred for large-volume resuscitation.

Metabolic Alkalosis and Contraction Alkalosis

This is the most common acid-base abnormality on a heart failure unit and the one the CMC exam is most likely to place in a cardiac vignette.

Mechanism. A loop diuretic removes sodium, chloride and water. The remaining bicarbonate is concentrated in a smaller extracellular volume (contraction alkalosis), while volume depletion activates aldosterone, which drives distal hydrogen and potassium secretion. Nasogastric suction and vomiting add direct loss of hydrochloric acid.

The signature: high pH, high bicarbonate, low potassium, low chloride, and a urine chloride below 20 mEq/L, which identifies the alkalosis as chloride-responsive.

Treatment is chloride and potassium repletion, usually normal saline with potassium chloride, which is precisely the therapy a congested heart failure patient may not tolerate. Alternatives include acetazolamide 250-500 mg IV or orally, which produces a bicarbonate diuresis, and adding a potassium-sparing agent such as spironolactone. Severe refractory alkalosis occasionally requires a hydrochloric acid infusion through a central line.

Why the nurse escalates it: alkalosis drives potassium intracellularly and worsens hypokalemia, lowers ionized calcium, prolongs the QT interval and sets up torsades de pointes, potentiates digoxin toxicity, and blunts respiratory drive in a patient being weaned from ventilatory support.

Cardiac Consequences of Acid-Base Derangement

Acidosis

Below a pH of about 7.20 the effects become clinically dominant:

• Reduced myocardial contractility and a lowered ventricular fibrillation threshold
• Blunted catecholamine and vasopressor responsiveness through adrenergic receptor desensitisation. The bedside presentation is the norepinephrine dose climbing with no pressure response — check the pH before adding a fourth vasopressor
• Hyperkalemia from extracellular potassium shift, worsening the arrhythmia risk
• Raised pulmonary vascular resistance and therefore right ventricular afterload. This is the reason acidosis is so dangerous in right ventricular infarction, pulmonary arterial hypertension, acute pulmonary embolism, and right ventricular failure after left ventricular assist device implantation

Sodium bicarbonate is not the treatment for acidosis from hypoperfusion. It generates carbon dioxide that a failing ventilatory pump must eliminate, can worsen intracellular acidosis, lowers ionized calcium, and delivers a sodium and osmolar load. Its evidence-supported roles are hyperkalemia, sodium-channel-blocker and tricyclic antidepressant toxicity, severe hyperchloremic acidosis, and selected cases of severe acidemia with hemodynamic instability, particularly with acute kidney injury. The treatment for a lactic acidosis of shock is restoring cardiac output and oxygen delivery.

Alkalosis

• Left shift of the oxyhemoglobin dissociation curve — hemoglobin binds oxygen more avidly and releases less at the tissue, so oxygen delivery falls even while the saturation looks excellent
• Hypokalemia and hypomagnesemia with QT prolongation and torsades
• Fall in ionized calcium as more calcium binds to albumin, producing hypotension, reduced contractility, tetany and paresthesias
• Cerebral and coronary vasoconstriction

Permissive Hypercapnia

Lung-protective ventilation accepts a PaCO2 of 60-80 mmHg and a pH of about 7.20-7.25 to keep tidal volumes at 4-8 mL/kg of predicted body weight and plateau pressure at 30 cm H2O or less. The cardiovascular cost is real: pulmonary vasoconstriction with increased right ventricular afterload, a catecholamine surge with tachyarrhythmia, and raised intracranial pressure. Permissive hypercapnia is therefore relatively contraindicated in right ventricular failure, pulmonary hypertension and raised intracranial pressure — precisely the populations a cardiac unit ventilates.

The Oxyhemoglobin Dissociation Curve

Normal P50, the PaO2 at which hemoglobin is 50% saturated, is about 26.6 mmHg.

Mnemonic: CADET, face Right — Carbon dioxide, Acid, 2,3-DPG, Exercise and Temperature all shift the curve to the right.

Sampling Technique: Where Nurses Create False Results

Modified Allen test. The traditional pre-cannulation check occludes both radial and ulnar arteries until the hand blanches, then releases the ulnar artery; colour should return in roughly 5 to 15 seconds. Know the technique, and know the controversy: the test has poor reproducibility and has not been shown to predict ischemic complications, and the transradial coronary intervention literature has largely abandoned it as a mandatory gate. Document per institutional policy, and rely on serial distal perfusion assessment after cannulation — colour, temperature, capillary refill, sensation, motor function and a pulse oximetry waveform on the ipsilateral thumb.

Arterial line sampling. Discard 3 to 6 mL of waste, or use a closed in-line reservoir and return the volume. Inadequate waste means the specimen is diluted with flush solution, producing a falsely low PaCO2 and bicarbonate, a falsely low potassium and hematocrit, and a prolonged aPTT or ACT if the flush is heparinised.

Air bubbles. Room air has a PO2 near 150-160 mmHg and essentially no carbon dioxide. A retained bubble equilibrates with the sample and drives the PaO2 toward 150, lowers the PaCO2 and raises the pH. Expel bubbles immediately and cap the syringe.

Delay before analysis. Leukocytes and erythrocytes continue to consume oxygen and produce carbon dioxide in the syringe. A specimen sitting at room temperature beyond 15 to 30 minutes shows a falling PaO2 and pH with a rising PaCO2, exaggerated in marked leukocytosis. Analyse within 15 minutes. Icing plastic syringes is now discouraged because oxygen diffuses through the plastic; prompt analysis is the better answer.

Heparin. Only the thin coating of a pre-heparinised syringe is needed. Excess liquid heparin dilutes the sample, falsely lowering PaCO2 and bicarbonate, and lithium-heparin preparations bind calcium and falsely lower the ionized calcium and potassium.

Venous versus arterial gases. A venous gas is adequate for pH and for trending carbon dioxide in most hemodynamically stable patients. Central venous pH runs roughly 0.03-0.05 units below arterial and PvCO2 about 4-6 mmHg above arterial, with peripheral venous samples running 3-8 mmHg higher. Two rules matter: a venous gas cannot assess oxygenation at all, and the venous-to-arterial relationship collapses in shock and cardiac arrest, when a venous sample reflects stagnant tissue rather than arterial physiology. A useful screening rule is that a venous PCO2 below 45 mmHg effectively excludes arterial hypercapnia.

Four-Disorder Reference Table