The Stewart-Figge acid-base model has been criticized for being mathematically complex. We aimed to develop simpler formalisms, which can be used at the bedside. The following simplifications were used: (1) [Ca2+] and [Mg2+] are replaced by their mid-reference concentrations (2) pH is set to 7.4. In the new model [SIDa] is replaced by its adjusted form, [SIDa, adj] = [Na+] + [K+] - [Cl-] + 6.5 and [SIG] is replaced by "bicarbonate gap", [BICgap] = [SIDa, adj] - (0.28⋅[Albumin]) - (1.82⋅[Phosphatei])- [HCO3̄]. The diagnostic performance of the model was tested in 210 patients with acute respiratory diseases and 17 healthy volunteers. [BICgap] was also compared to albumin-corrected anion gap ([AGc]). The concordant correlation coefficient between [SIDa, adj] and [SIDa] and between [BICgap] and [SIG] was 0.98 in both comparisons. The mean bias (limits of agreement) of [SIDa, adj] - [SIDa] and of [BICgap] - [SIG] were 0.53 meq/l (- 0.46 to 1.53) and 0.50 meq/l (- 0.70 to 1.70), respectively. A [SIDa, adj] < 50.4 meq/l had an accuracy of 0.995 (p < 0.001) for the diagnosis of strong ion (SI) acidosis, while a [SIDa, adj] > 52.5 meq/l had an accuracy of 0.997 (p < 0.001) for the diagnosis of SI alkalosis. A [BICgap] > 11.6 meq/l predicted unmeasured ion (UI) acidosis with an accuracy of 0.997 (p < 0.001), while an [AGc] > 19.88 meq/l predicted UI acidosis with an accuracy of 0.994 (p < 0.001). The "[BICgap] model" is a reliable tool for the assessment of acid-base disorders in patients with acute respiratory diseases. [BICgap] is not inferior to [AGc] in the diagnosis of UI acidosis.
Keywords: Anion gap; Base excess; Metabolic acidosis; Stewart-Figge acid–base model; Unmeasured ions.