Tales of the Anion Gap, Part III: Case Examples

Calculation and interpretation of the anion gap is extremely useful in the evaluation and treatment of the patient with metabolic acidosis. In this installment, we look at case examples involving patients with multiple sclerosis and urosepsis; COPD, diabetes, and renal failure; and type 1 diabetes.

As previously noted, calculation and interpretation of the anion gap is extremely useful in the evaluation and treatment of the patient with metabolic acidosis.

Recall that the evaluation of a patient with suspected metabolic acidosis should generally follow the following algorithm:

1. Is the patient acidotic?

2. Is the respiratory compensation appropriate?

3. Does the patient have an elevated or a normal anion gap acidosis?

4. If the anion gap is elevated, what is the unmeasured anion responsible?

Let’s work through a few case examples:

Case 1: A 47-year-old woman with late-stage multiple sclerosis is admitted with possible urosepsis. Her T is 38.9, pulse 112, BP 82/50. Her venous blood gases show pO2 42, pCO2 31, pH 7.33, and HCO3 18. Her chemistries show: Na+139, K 4.2, Cl 102, CO2 19, BUN 19, Cr 1.5, and albumin 3.6.

The patient’s pH and serum HCO3 are low, out of the normal range, so she is indeed acidotic. Using the formula for respiratory compensation as pCO2= (1.5 (serum bicarbonate) + 8), we see a calculated pCO2 of 35, so that is also appropriate. The anion gap is calculated as 139- (102+19) = 18, so this is indeed an elevated anion gap metabolic acidosis.

Recall that the unmeasured anion must be one or more of three categories: serum proteins, organic acids, or inorganic acids. In the current case, the patient’s serum L-lactate level came back 7.5mmoles/L, indicating that L-lactate is responsible for the majority of the patient’s elevated anion gap (ie, 7.5 out of roughly (18-6)=12). There are also probably some contributions to the anion gap by minor organic and inorganic acids (such as phosphates and sulfates) related to the patient’s renal dysfunction.

Case 2: A 61-year old man with COPD and long-term diabetes mellitus and diabetic kidney disease is admitted with worsening renal failure, fluid retention, and advancing renal failure. His venous blood gases are: pO2 39, pCO2 39, pH 7.28, and HCO3 18. The blood chemistries show: glucose 261, Na+132, K 5.4, Cl 102, CO2 18, and albumin 2.4.

Since the serum pH and HCO3 are less than the normal range, the patient is acidotic. With a serum HCO3 of 18, his estimated pCO2 after respiratory compensation should be about 1.5 (18)+ 8 = 35. Since his actual pCO2 is 39, he may also have a mild respiratory acidosis as well as his metabolic acidosis. His anion gap is calculated at 12, and if one were not aware that the normal anion gap is approximately 6 and that each gram/dl of serum albumin less than the normal range shrinks the anion gap by about 2.5, one might be tempted to call this a normal anion gap acidosis.

In this case, however, the ‘true’ anion gap would be approximately 16 versus a normal of 6 and therefore is certainly an elevated anion gap metabolic acidosis. Since plasma proteins are already accounted for in this calculation, the unmeasured anion must therefore be an organic or inorganic acid. The patient’s additional laboratory showed a serum lactate of 1.5 and negative serum ketones, so the unmeasured anions must mostly be inorganic acids such as sulfates and phosphates which accumulate in renal failure and minor organic acids which may also accumulate. This patient’s serum phosphorus level, in fact, was 8.1mg/dl, which is equivalent to 5.27Meq/L of inorganic acid and therefore accounts for most of the unmeasured anion in this case.

Case 3: A 21-year-old patient with Type 1 diabetes has been non-compliant with his insulin dosing and presents with fatigue, altered mental status, and Kussmaul respirations. His venous blood gases showed: pO2 45, pCO2 12, pH 7.01, HCO3 3. Blood chemistries showed glucose 734, Na+ 119, K 5.6, Cl 86, CO2 3, albumin 3.8, BUN 22, and Cr 1.7

The patient certainly has a metabolic acidosis. Our formula for predicted pCO2 related to respiratory compensation for metabolic acidosis is less useful at very low HCO3 levels as most all patients will find it impossible to hyperventilate to such a degree that the pCO2 would drop less than 12; the patient here is near maximal respiratory compensation for his metabolic acidosis. His calculated anion gap is 30; therefore, he certainly has an elevated anion gap metabolic acidosis. With serum proteins already accounted for, the unmeasured anions in this case must be organic or inorganic acids. In this patient’s case, his serum ketones were reported at 8.33mmoles/L and his serum L-lactate was 2.7mmoles/L.

In diabetic ketoacidosis, the lack of insulin causes the body to metabolize triglycerides and amino acids for energy, and with the aid of gluconeogenesis, forming free fatty acids that are metabolized in the mitochondria to ketones, which are strong organic acids. Lactic acidosis is also common in DKA, related to tissue hypoperfusion but also related to altered glucose metabolism. In this case, the measured ketones and lactate add up to only about 11Meq/L of the total anion gap of 30. The patient did have some degree of renal dysfunction, which might cause an increased level of phosphates, sulfates and other inorganic acids but not enough to account for the rest of the unmeasured anions.

The major acidic ketones are acetoacetate and beta-hydroxybutyrate (B-OH) and the ratio between these two acid anions primarily depends on the acidic state ie the NADH/NAD+ ratio. In severe DKA, the vast majority of the acidic ketone bodies, therefore, are usually B-OH, however this is not always the case and the concentration of acetoacetate may be significant. This is of importance in our calculations because acetoacetate is not measured on most clinical ketone assays; therefore the degree of ketoacidosis is usually underestimated-that is, a still unmeasured anion. By contrast, most urinary reagent strips do not measure B-OH but rather acetoacetate. Acetone is electrically neutral and does not enter into our calculations.

This patient was treated aggressively for his DKA and dehydration and received an insulin drip as well as several liters of crystalloid. A day later, his venous gases showed: pCO2 32, HCO3 15, pH 7.30, and his chemistries showed: glucose 132, Na+ 134, Cl 111, and total CO2 15.

Applying our algorithm, the patient is still acidotic, though significantly improved versus admission. His predicted pCO2 for respiratory compensation for the metabolic acidosis would be 30.5, so that is appropriate.

The anion gap is measured at 8, which is only slightly above normal. Therefore, the current picture would be called a normal anion gap (ie, hyperchloremic acidosis). Why is this common upon recovery from DKA?

Acidic ketone bodies are moderately strong organic acids and are excreted in the urine with a cation to maintain electrical neutrality. This is sometimes referred to as ‘loss of potential bicarbonate.’ To maintain electrical neutrality, HCO3- is generally replaced by Cl-, especially when the patient has received large amounts of normal saline in resuscitation. The effect is to cause a rise in plasma Cl- and the anion gap returns towards normal despite the persistence of the metabolic acidosis. Note that this is one of the uses of the delta ratio in evaluating metabolic acidosis, wherein the delta ratio is defined as the change in the anion gap divided by the fall in bicarbonate. In the current case, the patient’s delta ratio was roughly 24/22=1.1 on admission (typical of DKA and most elevated anion gap metabolic acidosis patients) but was roughly 2/10=0.2 the following day indicating a hyperchloremic state.

Delta ratios <0.4 are, by convention, suggestive of a hyperchloremic normal anion gap acidosis. DKA usually has a delta ratio of about 1 because of the previously mentioned urinary ketone loss. Lactic acidosis usually has a delta ratio in the 1-2 range with an average of about 1.6. A delta ratio>2 is indicative of a concomitant metabolic alkalosis or preexisting metabolic compensation for a respiratory acidosis.

In the next entry, we will review more complicated metabolic issues to illustrate the utility of these approaches in evaluation and management of the acidotic patient.