The serum anion gap is decreased in hyperchloremic (HCl) acidosis and
increased in diuretic-induced alkalosis, These anion gap changes have
been largely attributed to titration-induced variations in the net neg
ative charge of the serum proteins, which are the predominant non-HCO3
buffers of serum. It has recently been shown, however, that albumin h
as ail of the net protein charge, and titration-induced changes in cha
rge are smaller than have been widely believed, Because the non-HCO3 b
uffers are also titrated in acute hypocapnia and hypercapnia, these di
sorders were induced in 16 anesthetized dogs for 10 min in order to as
sess the effect of acute changes in pH on the anion gap, Although the
mean arterial pH varied from 7.04 to 7.65, the calculated mean albumin
charge only varied from 6.8 to 9.0 mEq/L. When the anion gap was comp
uted with HCO3 (AG(HCO3) = Na + K - Cl - HCO3), the change in AG(HCO3)
per 0.1 change in pH (Delta AG(HCO3)/ Delta pH) was only 0.15 mEq/L p
er 0.1 pH, When the anion gap was computed with total CO2 content (AG(
TCO2) = Na + K - CI - TCO2), Delta AG(TCO2)/Delta pH was larger (0.51
mEq/L per 0.1 pH) because of the effect of variable PCO2 levels on TCO
2, In a review of 22 previous studies in humans and dogs, similar esti
mates of Delta AG/Delta pH were obtained (after adjusting for the lowe
r albumin level in dogs), These results show that simple titration pro
cesses that occur within 10 min of a change in pH cause minimal change
s in the anion gap, Titration of the known non-HCO3 buffers of serum d
oes not explain the much larger anion gap changes of HCl acidosis and
diuretic alkalosis.