10.2 Acid-Base and Electrolyte Disorders

Every blood gas in a brain-injured patient is also a cerebral-blood-flow decision. This independent OpenExamPrep section covers anion-gap versus non-gap acidosis, Winters formula, metabolic alkalosis after diuresis or nasogastric suction, how respiratory acid-base moves ICP, and the electrolyte cluster that actually appears on neuro ICU rounds: sodium pattern recognition, potassium, magnesium, phosphate, and calcium after massive transfusion. Detailed syndrome of inappropriate antidiuretic hormone secretion (SIADH), cerebral salt wasting (CSW), and diabetes insipidus (DI) protocols live in the next chapter; you only need the pointers that keep you from treating the wrong water problem tonight.

Anion gap versus non-gap acidosis

The anion gap (AG) is Na+ − (Cl− + HCO3−). A typical albumin-normal gap is about 8–12 mEq/L; hypoalbuminemia shrinks the gap (roughly 2.5–3 mEq/L of gap per 1 g/dL of albumin lost), so a “normal” gap of 10 in a patient with albumin 1.8 g/dL can hide unmeasured anions. Always interpret the gap with albumin.

High-AG acidosis in this unit is usually lactate (seizure, ischemia, shock, propofol-related infusion syndrome, bowel catastrophe), ketoacids, toxic alcohols, or uremia. Normal-AG (hyperchloremic) acidosis is the everyday osmotherapy tax: 3% and 23.4% sodium chloride dump chloride. Large-volume 0.9% saline does the same. Carbonic anhydrase inhibitors, diarrhea, and renal tubular acidosis belong on the list but are less common than the chloride load you prescribed.

A chloride of 118 mEq/L after a night of 3% saline with bicarbonate 16 mEq/L and sodium 148 mEq/L is a non-gap story (AG = 148 − 118 − 16 = 14, near-normal if albumin is low). The same bicarbonate with chloride 100 mEq/L and sodium 140 mEq/L is a high gap (AG = 24) and needs a lactate and a toxin screen, not another saline bolus.

Worked Winters formula

In a metabolic acidosis, expected arterial PaCO2 ≈ 1.5 × [HCO3−] + 8 ± 2 (Winters). If measured PaCO2 sits in that window, respiratory compensation is appropriate. If PaCO2 is higher, there is a coexisting respiratory acidosis (the patient is not ventilating enough). If PaCO2 is lower, there is additional respiratory alkalosis (pain, anxiety, or over-ventilation).

Worked example. Bicarbonate is 12 mEq/L. Expected PaCO2 = 1.5 × 12 + 8 = 26 ± 2 mmHg, so 24–28 mmHg.

• Measured PaCO2 26 mmHg: compensation is on target.
• Measured PaCO2 38 mmHg: there are about 12 extra mmHg of CO2. That is a second disorder—airway obstruction, oversedation, high ICP with a suppressed respiratory drive, or a lung problem. Carbon dioxide dilates cerebral arterioles; that extra CO2 raises cerebral blood volume and ICP.
• Measured PaCO2 18 mmHg: extra hyperventilation. ICP may fall, but you risk cerebral ischemia, especially if you drive PaCO2 toward the mid-20s or below. Brain Trauma Foundation material treats brief hyperventilation as a bridge, not a lifestyle, and warns against PaCO2 <25 mmHg.

On the ventilator, do not “leave PaCO2 wherever it landed” after a bicarbonate of 10 mEq/L. Calculate Winters, then set minute ventilation on purpose. A patient with HTS-induced hyperchloremic acidosis who is still on a postoperative PaCO2 of 42 mmHg is acidemic and vasodilated for no neurologic benefit.

Metabolic alkalosis after diuresis and nasogastric losses

Loop and thiazide diuretics contract chloride and generate bicarbonate. Nasogastric (NG) suction or high gastric output removes hydrochloric acid. The result is a chloride-responsive metabolic alkalosis: high bicarbonate, high PaCO2 as compensation, often hypokalemia, and a urine chloride that is low once the diuretic has worn off. Volume restoration with chloride (usually 0.9% saline if the brain sodium target allows, or potassium chloride if potassium is the limiter) shuts off the alkalosis. Acetazolamide is sometimes used to dump bicarbonate when volume is already high, but it can drop sodium and worsen ICP physiology if you are not watching osmoles.

Compensation for metabolic alkalosis is hypoventilation. That raises PaCO2 and can raise ICP. In a patient with a fresh hemicraniectomy you may accept a bit of alkalosis rather than hypoventilate them into hypercapnia. In a closed skull with ICP 25 mmHg, you treat the alkalosis and keep PaCO2 in a planned range, often near 35–40 mmHg unless you are bridging a herniation event.

Citrate from blood products or CRRT is metabolized to bicarbonate and can add a metabolic alkalosis once shock settles. That is a different timeline from NG suction.

Respiratory acid-base versus ICP

Recall the directional rules:

• Hypercapnia → cerebral vasodilation → higher ICP
• Hypocapnia → vasoconstriction → lower ICP, at the cost of ischemia if extreme or prolonged
• Hypoxemia also dilates cerebral vessels

Metabolic acidemia plus a high PaCO2 is a double hit. Metabolic alkalosis plus a compensatory high PaCO2 is a quieter ICP risk that people miss because the pH looks “almost normal.” Always read PaCO2 against the neurologic exam and the ICP number, not against a medical-ICU habit of permitting permissive hypercapnia in ARDS without a brain caveat. If the lungs need a higher PaCO2 strategy, you need an ICP plan: drainage, osmotherapy, sedation, or surgery—not hope.

Sodium pointers: SIADH versus CSW versus DI

Full diagnostic algorithms and treatment (fluid restriction versus volume, fludrocortisone, desmopressin, 3% saline) belong in the endocrine and salt-water chapter. For this section, lock the pattern:

The dangerous error is treating CSW with fluid restriction as if it were SIADH. You dry a vasospastic SAH patient and invite delayed cerebral ischemia. The other dangerous error is treating DI with free-water restriction because “sodium is high.” You will cook the remaining neurons. If volume status is ambiguous, examine the patient, look at input/output, hematocrit, and urea, and do not start a water-restriction order set on a subarachnoid-hemorrhage night shift. Chapter 11 carries the treatment detail.

Potassium, magnesium, and phosphate

Potassium falls with mannitol kaliuresis, beta-agonists, insulin, alkalosis, and poor intake. It rises with succinylcholine in denervated or burned muscle, acidosis, rhabdomyolysis, and missed dialysis. Hypokalemia promotes ventricular ectopy and can weaken the diaphragm. Replace while you fix the driver. Remember that hypomagnesemia makes hypokalemia refractory; give both.

Magnesium is a cofactor for potassium retention and a membrane stabilizer. Levels drop with osmotherapy, diuretics, ethanol, and refeeding. Clinically you see tremor, tetany, arrhythmias, and seizures. In aneurysmal SAH, magnesium is sometimes discussed for vasospasm physiology, but you still treat a low level as a seizure and arrhythmia risk rather than as a proven outcome drug. Target a solidly normal magnesium when the patient is spasm-prone or on QT-prolonging meds.

Phosphate falls with refeeding: carbohydrate and insulin drive phosphate (and potassium and magnesium) indoors after starvation, alcohol use, or a long NPO stretch. Osmotherapy and continuous CRRT also strip phosphate. Severe hypophosphatemia causes weakness, failure to wean, hemolysis, and seizures. Check phosphate before starting full-rate feeding in a depleted patient, and replace aggressively rather than celebrating a climbing prealbumin.

Calcium after massive transfusion

Stored red cells and plasma contain citrate, which chelates calcium. A slow transfusion in a patient with a working liver is usually tolerated. Massive transfusion, liver failure, hypothermia, and citrate CRRT overwhelm citrate metabolism. Total calcium bound to albumin can look acceptable while ionized calcium is 0.7 mmol/L and the patient is hypotensive and coagulopathic. Treat ionized hypocalcemia with intravenous calcium, warm the patient, and slow citrate if the circuit is the source. Do not use a corrected total calcium formula as your only number during a massive transfusion.

Hypercalcemia is less common here but appears with immobilization, thiazides, and some malignancies. It causes polyuria that can mimic DI until you look at the calcium.

Putting a gas and a panel together

Order of attack that matches how items are written:

1. Look at pH, then PaCO2 and bicarbonate, then AG (albumin-adjusted).
2. If metabolic acidosis, run Winters and ask whether the ventilator is helping or hurting ICP.
3. If chloride is the story, look at HTS and saline volume.
4. If sodium is abnormal, sort volume and urine osmolality before you restrict or free-water-load.
5. Replace K, Mg, and phosphate on the same round you start nutrition.
6. If blood is pouring in, check ionized calcium rather than adding another amp of bicarbonate “for the pH.”

Exam traps

Calling every low bicarbonate “lactic acidosis” when chloride is 120 mEq/L after 23.4% saline. Ignoring a PaCO2 of 40 mmHg when Winters predicted 26 mmHg. Fluid-restricting a dry SAH patient because urine sodium is high. Starting tube feeds at goal in a cachectic patient without checking phosphate. Transfusing 8 units and chasing MAP with vasopressors while ionized calcium is 0.72 mmol/L.