High Anion Gap Metabolic Acidosis (HAGMA)

Calculate the corrected gap, then the delta-delta - Any anion gap over 20 is a primary metabolic acidosis regardless of what the pH and bicarbonate look like. Once a gap is present, check the delta-delta even if the pH is normal: delta-gap added to the bicarbonate above 30 means a concurrent alkalosis, below 23 means a second acidosis. Mixed disorders are the rule in these patients, not the exception.

Screen for the cause with a defined panel - Osmolar gap (measured minus calculated osmolarity, normal 10 or less), lactate, ketones, salicylate, acetaminophen, creatinine, blood gas, and urinalysis. An elevated osmolar gap points at a toxic alcohol; calcium oxalate crystals on urine microscopy point specifically at ethylene glycol. An osmolar gap is expected in ketoacidosis because acetone raises osmolarity, so trend it: failure to fall with treatment should raise toxic alcohol.

D-lactate in the short-gut patient - After gastric bypass or in short gut syndrome, colonic bacteria produce D-lactate, which the standard serum lactate assay does not detect. A gap acidosis with a normal lactate in that patient needs the send-out.

Contractility falls before you see it - Acidemia begins degrading myocardial contractility around pH 7.2, initially masked by the catecholamine surge that acidemia itself provokes. Below 7.1 the heart becomes unresponsive to catecholamines and decompensation is abrupt. Have vasopressor and inotrope drawn before induction in anyone arriving with a pH under 7.2.

Rising pressor requirement is the acid, not the pressor - Lactic acidosis reduces cardiac contractility and produces vascular hyporesponsiveness to vasopressors. Escalating the infusion without addressing perfusion and the underlying cause chases a moving target.

Do not take away the respiratory compensation - These patients are already hyperventilating, often with Kussmaul breathing; expected pCO2 = 1.5 x [HCO3-] + 8, bottoming out around 15 mm Hg as bicarbonate falls from 10 to 5. Induction, paralysis, and default ventilator settings can deliver less minute ventilation than they were generating themselves. Match their pre-induction minute ventilation, keep the apneic period short, and check a gas early.

A respiratory component is worse than a metabolic one - CO2 crosses cell membranes far more readily than fixed acid, so hypoventilation on top of an existing metabolic acidosis drops intracellular pH fast. Avoid rebreathing and avoid letting them ride a high end-tidal CO2 on emergence.

Bicarbonate is reserved, not routine - Alkalinization is for severe acidosis, pH under 7.1, and under 6.9 in diabetic ketoacidosis. The mechanism is worth knowing: sodium bicarbonate raises pH because it has a positive strong ion difference, not because of the bicarbonate ion itself. By the same logic, resuscitating with 0.9% saline, which has a strong ion difference of zero, will deepen an existing acidosis. Reach for a balanced solution.

Potassium moves with the pH - Acidemia shifts potassium out of cells, so the measured serum value overstates total body potassium and will fall as you correct the acidosis. This bites hardest in DKA, where osmotic diuresis has already stripped total body potassium and sodium. Recheck after treatment starts, not just before it.

Type A versus type B lactate changes what you do - Type A is hypoperfusion: all shock states, regional or mesenteric ischemia, seizures, severe shivering. It responds to restoring perfusion. Type B has nothing to do with tissue hypoxia (liver disease, malignancy, epinephrine, TPN, thiamine deficiency, mitochondrial myopathy, DKA, ethanol) and more fluid or more pressor will not fix it. Remember your own epinephrine infusion is on that list.

Treat the cause, and know what is not the cause - DKA needs insulin and glucose; sepsis needs broad-spectrum antibiotics within an hour of recognition, source control, and 30 mL/kg crystalloid; toxic alcohol needs fomepizole, most effective within 6 hours of presentation, with emergent hemodialysis prepared for severe recalcitrant acidosis. Metformin is not a cause; a Cochrane review found substantial evidence against metformin-induced lactic acidosis. Phenformin was the real offender.

Know the numbers that predict trouble - Lactate under 2 mmol/L is normal, 2 to 4 is hyperlactatemia, 4 or above is severe. Shock with severe lactic acidosis (pH under 7.2) carries roughly 50% mortality, and no survival has been reported in shock once pH falls below 7.0. How fast the lactate clears tracks survival, so trend it rather than treating a single value.

A hand vein gas will usually do - Venous blood from the dorsum of the hand is partly arterialized under general anesthesia; pCO2 runs about 5 mm Hg off arterial and pH by 0.03 to 0.04 units. Air bubbles, heparin (which is acidic), and leukocyte larceny all corrupt the sample, so cool it and run it promptly.

A high anion gap metabolic acidosis means an unmeasured anion, such as lactate, ketoacids, salicylate, the metabolites of methanol and ethylene glycol, or uremic anions, has consumed bicarbonate. Anion gap = Na - (Cl + HCO3), traditionally normal at 8 to 12 mEq/L, and it must be corrected for albumin: add 2.5 mEq/L for every 1 g/dL the albumin sits below 4. Skip that correction in a hypoalbuminemic ICU patient and a real gap acidosis reads as normal.

Separating gap from non-gap acidosis is what points at the etiology, and the etiology is what you treat. The differential is carried by the GOLD MARK and CATMUDPILES mnemonics; in practice it comes down to lactate, ketoacidosis, renal failure, salicylate, toxic alcohols, and massive rhabdomyolysis. What makes it matter in the OR is the acidemia itself: below a pH of 7.2 contractility falls, and below 7.1 the myocardium stops answering catecholamines.