When a blood-gas analyser is not available, ionised calcium often has to be estimated from a routine total calcium, the serum albumin, and the patient’s pH. This tool chains the three standard adjustments so you get a defensible estimate rather than a raw total that may be misleading.
How it works
Three steps are combined:
corrected_Ca = total_Ca + 0.02 × (40 − albumin_g_L) (Payne)
iCa_raw = 0.5 × corrected_Ca (≈50% ionised)
iCa_pH = iCa_raw − 0.05 × ((pH − 7.40) / 0.1) (pH correction)The Payne step removes the effect of low albumin on the total. Roughly half of the albumin-corrected total is ionised at pH 7.40. The pH step then nudges the ionised value up in acidosis and down in alkalosis, because calcium competes with hydrogen ions for albumin binding.
Why total calcium alone is misleading in sick patients
Total serum calcium has three fractions:
- Protein-bound calcium (~40%) — mostly bound to albumin, not physiologically active. Fluctuates with albumin level.
- Complex-bound calcium (~10%) — bound to anions such as bicarbonate, phosphate, and citrate. Not active.
- Free ionised calcium (~50%) — the physiologically active fraction that drives nerve and muscle function, clotting, cardiac action, and hormone release.
Only ionised calcium matters clinically. Total calcium is the routine laboratory measurement because it is simpler to measure, but in any patient with low albumin (critically ill patients, liver disease, nephrotic syndrome, malnutrition) the total calcium substantially underestimates the ionised fraction. Treating a “low” total calcium in a hypoalbuminaemic patient with calcium replacement that is not needed is a preventable error.
The Payne albumin correction
The Payne formula corrects total calcium to the equivalent at a normal albumin of 40 g/L:
Corrected Ca (mmol/L) = total Ca + 0.02 × (40 − albumin in g/L)For each g/L albumin falls below 40, the corrected calcium rises by 0.02 mmol/L, accounting for the smaller protein-bound fraction. In mg/dL the equivalent coefficient is 0.8 per g/dL below the normal albumin of 4.0 g/dL.
This correction is widely used and taught, but it has well-recognised limitations. It performs best near the population mean from which the regression was derived and less well at extremes of albumin or calcium. It also cannot account for variation in binding affinity from acidosis, drugs, or abnormal protein fractions.
pH correction: acidosis and alkalosis
Hydrogen ions and calcium ions compete for the same binding sites on albumin. In acidosis, the excess H⁺ displaces calcium from albumin, increasing the ionised fraction. In alkalosis, more calcium binds to albumin, lowering the ionised fraction. The classical correction is approximately 0.05 mmol/L change in ionised calcium per 0.1 pH unit, in the opposite direction to pH.
For example, a patient with pH 7.20 (acidosis, 0.20 below 7.40) would have an estimated ionised calcium roughly 0.10 mmol/L higher than at pH 7.40. Conversely, a patient hyperventilating to pH 7.55 would have an estimated ionised calcium 0.075 mmol/L lower.
This correction matters because acute acid-base disorders are common in critical care, and the pH effect on ionised calcium is clinically significant.
Practical worked example
A patient has total calcium 2.10 mmol/L, albumin 22 g/L, pH 7.28.
- Payne correction:
2.10 + 0.02 × (40 − 22) = 2.10 + 0.36 = 2.46 mmol/L - Ionised raw:
0.5 × 2.46 = 1.23 mmol/L - pH correction:
1.23 − 0.05 × ((7.28 − 7.40) / 0.1) = 1.23 − 0.05 × (−1.2) = 1.23 + 0.06 = 1.29 mmol/L
Estimated ionised calcium: ~1.29 mmol/L — within the normal range despite the low total calcium. The low albumin had masked what is, on this estimate, a borderline-normal ionised calcium. A directly measured ionised calcium on a blood gas would confirm.
Example and notes
A total calcium of 2.40 mmol/L with albumin 28 g/L and pH 7.40 corrects to about 2.64 mmol/L, giving an estimated ionised calcium near 1.32 mmol/L — higher than the raw total would suggest, because the low albumin had masked it. Treat the result as a guide only: a directly measured ionised calcium on a fresh, anaerobic sample is always preferred, and all values should be read alongside the clinical picture and trend. This tool is intended to support, not replace, clinical judgement.