Normal AG Values

As shown in Table 4.3, the normal reference range for AG is 12 ± 4. This value was derived from older methods of electrolyte determinations using colorimetry and flame photometry. With the introduction of ion-selective electrode methodology, the normal AG levels are much lower, ranging from 3 to 11. These low values are related to high Cl⁻ determination. Variations in normal values of AG are frequently seen from laboratory to laboratory, and the clinician should follow his or her laboratory values for proper interpretation of AG in clinical medicine. For simplicity, an AG of 10 mEq/L is considered normal.

Hyperglycemia and AG

It is always debated whether or not AG should be calculated using measured serum Na⁺ or corrected Na⁺ in a patient with severe hyperglycemia. It is suggested that only measured Na⁺ should be used for AG, and not the corrected Na⁺. This suggestion is based on the assumption that Cl⁻ and HCO₃ ⁻ are equally diluted as Na⁺ by movement of water from inside to outside of the cell caused by hyperglycemia. When AG is calculated using only corrected Na⁺ and not corrected Cl⁻ and HCO₃ ⁻ for hyperglycemia, the AG is overestimated with the implication of underlying high AG metabolic acidosis. Na⁺ decreases by 1.6 mEq/dL for each increase in 100 mg/dL glucose above normal glucose levels, but no such number is given for either Cl⁻ or HCO₃ ⁻. It is, thus, emphasized that only measured Na⁺ should be used to calculate the serum AG.

Clinical Use of AG

Traditionally, the AG is useful in classifying metabolic acidosis into high, normal, or low AG metabolic acidosis. High AG is due to accumulation of unmeasured anions, whereas normal AG is usually related to high Cl⁻ level. Low AG implies a substantial decrease in unmeasured anions, an increase in unmeasured cations, and spurious decrease in [Na⁺] or spurious increases in [Cl⁻] or [HCO₃ ⁻].

Mnemonic for High AG Metabolic Acidosis

Normal AG Metabolic Acidosis

Low AG Metabolic Acidosis and Correction for Low Serum Albumin

In addition to a decrease in unmeasured anions, an increase in unmeasured cations causes low AG. For example, A patient with IgG myeloma will have low AG because IgG molecules carry a positive charge at a pH of 7.4. Also, ingestion of bromide or iodine-containing medications or intoxication of these halides can raise the concentration of unmeasured anions and lower AG due to measurements of these halides as Cl⁻. Salicylate overdose usually gives a negative AG, as high salicylate levels are measured as high Cl⁻ levels by certain chloride-sensitive ion-selective electrodes.

Occasionally, patients with severe hypertriglyceridemia may present with low AG due to a different laboratory measurement. Severe hypercalcemia, hypermagnesemia, or lithium toxicity may cause low AG. Some case reports demonstrated low AG in hypotonic hyponatremic patients. Thus, low AG is caused by many conditions.

Use of ∆AG/∆HCO₃ ⁻

Not only is AG useful in the classification of metabolic acidosis, but it is also indirectly helpful in analyzing mixed acid–base disorders such as high AG metabolic acidosis and metabolic alkalosis or high AG and normal AG metabolic acidosis. In a simple or uncomplicated high AG metabolic acidosis, the increase in AG above normal (called ∆AG) is equal to the decrease in HCO₃ ⁻ from normal value (called ∆HCO₃). In other words, for every 1 mEq/L rise in the AG, there is a concomitant fall of 1 mEq/L in HCO₃ ⁻. This is defined as ∆AG/∆HCO₃ ⁻. In a simple high AG metabolic acidosis, the ∆AG/∆HCO₃ ⁻ ratio is 1. Any significant deviation from 1 is indicative of a mixed acid–base disorder. For example, A patient with diarrhea typically develops a hyperchloremic or normal AG metabolic acidosis. In this condition, the decrease in HCO₃ ⁻ results in a reciprocal increase in Cl⁻ so that the AG does not change. As a result, the ∆AG/∆HCO₃ ⁻ ratio is 0. If this patient develops hypotension and subsequent lactic acidosis, the HCO₃ ⁻ decreases even further. In other words, the ∆HCO₃ ⁻ is greater than ∆AG, and the ∆AG/∆HCO₃ ⁻ ratio is <1 (above 0 but below 1), indicating a mixed high and normal AG metabolic acidosis.

In pure lactic acidosis, the ∆AG/∆HCO₃ ⁻ ratio is usually 1.6 because the lactate anions remain in the extracellular compartment because of low urinary excretion. This raises the AG. However, the HCO₃ ⁻ concentration does not decrease as non-HCO₃ ⁻ buffers participate in buffering lactate anions. In my opinion, the ∆AG/∆HCO₃ ⁻ ratio is usually not helpful in pure lactic acidosis, as lactate levels are available, and the ∆AG/∆HCO₃ ⁻ ratio varies from 0.8 to 1.8.

In a mixed high AG metabolic acidosis and metabolic alkalosis, the HCO₃ ⁻ level is inappropriately high relative to the increase in AG. As a result, the ∆AG/∆HCO₃ ⁻ ratio is >2.

Although the ∆AG/∆HCO₃ ⁻ ratio is a useful tool, it should not be used alone in identifying mixed acid–base disorders. Other pieces of evidence such as clinical information about the patient, normal AG range, AG corrected for albumin, and uncovering of hidden acid–base disorder during treatment should be taken into account whenever a mixed acid–base disorder is analyzed. Also, evaluation of volume status is important, as total HCO₃ ⁻ content in the ECF compartment may change. An example is diabetic ketoacidosis (DKA). Prior to development of DKA, the ECF water content in a 70 kg patient is 14 L, and the total HCO₃ ⁻ content is 336 mEq (14 L × 24 mEq/L = 336 mEq). DKA obligates water loss, and the patient develops volume depletion. As a result, water content decreases presumably to 10 L. At the same time, the HCO₃ ⁻ content decreases due to buffering of ketoacids presumably to 15 mEq/L. Now the new total HCO₃ ⁻ content is 150 mEq (10 L × 15 mEq = 150), a deficit of 186 mEq (336 mEq – 150 mEq = 186 mEq). Note that the AG continues to increase because of the addition of ketoacids, but the concentration of HCO₃ ⁻ decreases. As a result, the ∆AG/∆HCO₃ ⁻ ratio will be >2, suggesting the presence of spurious metabolic alkalosis. Thus, one must be cautious in interpreting the ∆AG/∆HCO₃ ⁻ ratio.