Acid–base interpreter

The sequence follows Berend, de Vries & Gans (NEJM 2014): check the numbers, name the primary disorder, test whether the secondary response is what the primary disorder predicts, then use the anion gap, delta ratio and osmolar gap to uncover hidden metabolic processes.

0 · Internal consistency. [H⁺] (nmol/L) = 24 × PCO₂ ÷ HCO₃⁻ (Henderson), compared with [H⁺] from pH = 10^{(9 − pH)}. The tool flags a difference of more than 10% (≈ 0.04 pH units). This tolerance is a tool convention, not a published cut-off: it allows for rounding of reported values and for serum total CO₂ running about 0.03 × PCO₂ (≈ 1.2 mmol/L) above true bicarbonate. Blood-gas analysers calculate HCO₃⁻ from pH and PCO₂, so the check mainly catches transcription errors or a BMP drawn at a different time.

1 · pH. Acidemia < 7.35, alkalemia > 7.45. When pH is in range but PCO₂ and HCO₃⁻ are both abnormal, the side of 7.40 picks the likely primary process.

2 · Primary disorder. HCO₃⁻ outside normal ± 2 mmol/L or PaCO₂ outside normal ± 2 mmHg (Berend reference values 24 ± 2 and 40 ± 2), on the side that explains the pH. Both on the same side = two primary disorders.

PaCO₂ above the expected band = added respiratory acidosis; below = added respiratory alkalosis. For respiratory disorders, HCO₃⁻ below the acute prediction = added metabolic acidosis; above the chronic prediction = added metabolic alkalosis. The tool applies ± 2 mmol/L around the respiratory predictions (the HCO₃⁻ reference range). The 40 and 24 anchors follow the "normal" values you set.

Berend notes these equations come from decades-old human and canine studies; confidence limits for severe chronic disorders are uncertain, and only chronic respiratory alkalosis is generally accepted to normalise pH.

Figge 1998. Reported normal ranges vary from 3–12 to 8.5–15 mmol/L by analyser (Berend) — set the normal value to your lab's. An AG > 10 above the upper reference limit is highly suggestive of an organic acidosis; small rises are less specific.

Berend frames the same idea as ΔAG − ΔHCO₃⁻ = 0 ± 5 (or 0.6 × ΔAG − ΔHCO₃⁻ in lactic acidosis): > +5 suggests added metabolic alkalosis, < −5 added normal-gap acidosis.

Osmolar gap = measured − calculated; < 10 mOsm/kg is usually considered normal, but the population range is wide (−10 to +10), so a normal gap does not exclude toxic alcohol ingestion, and lactic acidosis or ketoacidosis can raise it (Berend). Ethanol factor from Purssell 2001.

In normal-gap acidosis a negative value indicates appropriate ammonium excretion (diarrhoea, proximal RTA); a positive value indicates impaired ammonium excretion (distal RTA, type 4 RTA/hypoaldosteronism, kidney failure) (Batlle 1988). Unreliable with urine Na⁺ < 20 mmol/L, urine pH > 6.5, polyuria, or unmeasured urinary anions such as ketoacids, hippurate or penicillins (Berend).

Peripheral venous pH runs about 0.03 lower than arterial and agrees well; venous PCO₂ does not — the 95% prediction interval of the venous–arterial difference spans roughly −10.7 to +2.4 mmHg (Byrne 2014 meta-analysis). Compensation rules were derived from arterial values, so treat any "mixed respiratory" conclusion from a VBG as provisional. Berend's venous reference values: pH 7.36–7.38, PvCO₂ 43–48 mmHg, HCO₃⁻ 25–26 mmol/L.