I teach this concept as determining what the bicarbonae would be in the absence of or prior to the anion gap.
The concept comes from the idea that for every mEq of bicarbonate that is consumed by the strong acid (other anion) the anion gap should rise by one. So if the bicarb is 16, a delta of 8, we would expect an anion gap of 20, a normal anion gap of 12 plus the delta bicarbonate of 8. This is a ∆AG/∆Bicarb of one.
If the patient had a pre-existing metabolic alkalosis with a bicarbonate of 30, then the patient would have a bicarbonate of 22 and an anion gap of 20. This would give ∆AG/∆Bicarb of 8/2 or 4.
If the patient had a pre-existing metabolic acidosis (non-anion gap) with a bicarbonate of 16, then the patient would have a bicarbonate of 8 and an anion gap of 20. This would give ∆AG/∆Bicarb of 8/16 or 0.5.
Concurrent metabolic alkalosis leads to ratios over 1 and preexisting metabolic acidosis (non-anion gap) yield a ratio below 1.
I had always been suspicious of this because the assumption of the one for one change in anion gap and bicarbonate. This didn't seem to be very biologic. Turns out my suspicion was justified as numerous studies (Androgue, Elisaf) have shown that the ratio does not hold up.
In this paper by Paulson et al they found:
[Some authors] suggested that mixed disturbances should be considered if the ratio is less than 0.8 or greater than 1.2. Paulson, applying this rule to a group of normal control subjects and patients with simple metabolic acidosis, noted that the formula erroneously categorized 56% [specificity of 44%] of this group as mixed disturbances. Use of the 95% confidence interval of ±8 mEq/L increased the specificity to 97% but with a poor sensitivity of only 27%.
That's terrible. Why torture the brains of medical students with this type of worthlessness.
Good review here.