Why total calcium needs correcting at all
Calcium circulates in three fractions. About half is free ionised calcium, the only fraction with physiological activity — it drives neuromuscular excitability, cardiac conduction, coagulation and hormone secretion. Roughly 40% is bound to protein, overwhelmingly albumin. The remaining tenth is complexed to small anions such as citrate, phosphate and bicarbonate.
A routine chemistry panel reports the sum of all three. That total therefore moves with albumin whether or not the ionised fraction has moved at all. Drop a patient's albumin from 4.0 to 2.0 g/dL — an entirely ordinary consequence of cirrhosis, nephrotic syndrome, sepsis or a long hospital stay — and the total calcium falls by about 1.6 mg/dL while the ionised calcium is unchanged. Read at face value, that patient looks hypocalcaemic and is not.
The correction rewinds that effect. It answers a narrow question: what would this patient's total calcium read if their albumin were normal? The answer is a better proxy for ionised calcium than the raw total, which is why it appears on nearly every inpatient chemistry report in some form. It is not a measurement of ionised calcium, and where the decision matters it does not replace one.
The Payne formula and where its numbers come from
Payne and colleagues regressed total calcium against albumin in a large hospital population and found a slope close to 0.02 mmol/L of calcium per 1 g/L of albumin. Converted to conventional units — 1 mmol/L of calcium is 4.008 mg/dL, and 1 g/L of albumin is 0.1 g/dL — that slope becomes 0.8 mg/dL of calcium per 1 g/dL of albumin. The two published forms of the formula are the same line in different units, not two different corrections.
The correction has exactly three moving parts:
- The measured calcium, which sets the starting point.
- The albumin difference, reference minus measured. This is the only term that can change sign. Below the reference it is positive and the correction adds; above the reference it is negative and the correction subtracts. A patient who is dehydrated with an albumin of 5.2 g/dL has their calcium adjusted down, and forgetting that is a real way to over-call hypercalcaemia.
- The coefficient, 0.8 mg/dL per g/dL by default. Some laboratories publish their own, derived from their own population and their own calcium and albumin methods. Bromocresol green and bromocresol purple albumin assays do not agree with each other, and a correction derived on one does not transfer cleanly to the other.
The reference albumin is the third convention worth checking. Payne used 4.0 g/dL (40 g/L), and that is this calculator's default, but a laboratory serving a population with a different mean albumin may normalise to 4.4 g/dL instead. Change the field if yours does; a 0.4 g/dL difference in the reference moves every corrected result by 0.32 mg/dL.
Worked example: a cirrhotic patient with a calcium of 8.0 mg/dL
A 58-year-old with alcohol-related cirrhosis has a total calcium of 8.0 mg/dL and an albumin of 2.0 g/dL. The panel flags the calcium as low. Work the correction by hand.
- Albumin difference. 4.0 − 2.0 = 2.0 g/dL below the reference.
- Adjustment. 0.8 × 2.0 = +1.6 mg/dL.
- Corrected calcium. 8.0 + 1.6 = 9.6 mg/dL.
- Convert to SI. 9.6 ÷ 4.008 = 2.395 mmol/L.
- Classify. Against a typical interval of 8.5 to 10.2 mg/dL (2.12 to 2.55 mmol/L), 9.6 sits comfortably inside it.
The measured value said hypocalcaemia; the corrected value says normal. Nothing about this patient's calcium homeostasis is disturbed — their albumin is. Chasing the raw number here leads to unnecessary calcium supplementation, repeat testing, and occasionally a parathyroid hormone assay that answers a question nobody needed to ask.
Run the same patient with an albumin of 4.0 g/dL and the adjustment is zero: measured and corrected coincide. That is the useful sanity check on any implementation — at the reference albumin the formula must do nothing at all.
How to read the corrected value
Most laboratories quote a total calcium interval of about 8.5 to 10.2 mg/dL (2.12 to 2.55 mmol/L), and the corrected value is compared against the same interval — that is the whole point of normalising to a reference albumin. The number to watch is not the corrected calcium alone but whether the correction moved it across a boundary. If measured and corrected land on the same side of the interval, the albumin was not doing much work. If they land on opposite sides, the albumin was doing all of it, and you should be reaching for an ionised calcium before acting.
A corrected calcium above the interval sends you to parathyroid hormone. Primary hyperparathyroidism and malignancy together account for the large majority of hypercalcaemia; a PTH that is high or inappropriately normal points to the former, a suppressed PTH to the latter. Above 14 mg/dL (3.5 mmol/L) the picture is an emergency regardless of cause.
A corrected calcium below the interval sends you to magnesium first. Hypomagnesaemia both impairs parathyroid hormone secretion and blunts its effect at the bone, and calcium given without correcting magnesium will not stay corrected. Then look at phosphate, vitamin D, renal function, and the surgical and transfusion history. Because hypocalcaemia prolongs the QT interval, a corrected QT is worth having before treatment in anyone symptomatic.
The correction's own accuracy deserves scepticism. Validation studies in dialysis and critical care populations have repeatedly found that albumin-corrected calcium agrees only moderately with directly measured ionised calcium, and misclassifies a meaningful minority of patients in both directions. Treat it as a triage tool that improves on the raw total, not as a substitute for the measurement.
How much the correction moves the number
| Albumin (g/dL) | Albumin (g/L) | Adjustment (mg/dL) | Adjustment (mmol/L) |
|---|---|---|---|
| 1.5 | 15 | +2.00 | +0.50 |
| 2.0 | 20 | +1.60 | +0.40 |
| 2.5 | 25 | +1.20 | +0.30 |
| 3.0 | 30 | +0.80 | +0.20 |
| 3.5 | 35 | +0.40 | +0.10 |
| 4.0 | 40 | 0.00 | 0.00 |
| 4.5 | 45 | −0.40 | −0.10 |
| 5.0 | 50 | −0.80 | −0.20 |
Each adjustment is 0.8 × (4.0 − albumin) in mg/dL, and the SI column is that figure divided by 4.008 and rounded to two decimals.
Pitfalls that make a corrected calcium wrong
- Using an albumin from a different day. Albumin falls fast with fluid resuscitation and inflammation. Pair the calcium with an albumin from the same tube.
- Applying the correction to an ionised calcium. The correction exists to approximate ionised calcium from a total. Applied to a measured ionised result it is meaningless.
- Forgetting the sign when albumin is high. Dehydration and prolonged tourniquet time raise albumin, and the correction then subtracts. A corrected value can be normal when the measured one looked high.
- Ignoring pH. Alkalosis increases calcium binding to albumin and lowers ionised calcium without changing the total, so the corrected value looks reassuring in a hyperventilating patient with tetany. Only an ionised calcium shows this.
- Ignoring citrate and paraproteins. Massive transfusion, regional citrate anticoagulation in dialysis, and myeloma paraproteins all shift calcium binding in ways the albumin term cannot see.
- Trusting it at very low albumin. Below about 2 g/dL the correction is extrapolating beyond its derivation population and performs poorly against measured ionised calcium.
- Mixing units. Entering albumin in g/L into a formula expecting g/dL inflates the correction roughly tenfold. This calculator converts for you — check the unit selector matches your report.
When to measure ionised calcium instead
Measure ionised calcium directly, and do not rely on any correction, when the patient is critically ill, on renal replacement therapy with citrate, receiving massive transfusion, markedly acidotic or alkalotic, has a paraproteinaemia, or when the corrected and measured values disagree about the diagnosis. Ionised calcium is measured on a blood gas analyser, needs an anaerobic sample handled promptly, and is reported directly in mmol/L with a typical interval near 1.15 to 1.33 mmol/L — check your own analyser's interval.
Where this sits among the other panel corrections
Corrected calcium belongs to a family of bedside adjustments that all do the same job: undo a predictable analytical or physiological distortion so a routine number can be read against its usual interval. The albumin-corrected anion gap corrects for exactly the same protein, in the opposite direction — low albumin shrinks the gap and lowers the total calcium at once, so a hypoalbuminaemic patient often needs both corrections at the same time. The glucose-corrected sodium does the same for dilutional hyponatraemia in hyperglycaemia.
Alongside these sit the normalisations that let a result be compared between patients rather than within one: eGFR normalises creatinine clearance to a standard body surface area, and Cockcroft-Gault clearance deliberately does not, because drug dosing needs the patient's absolute clearance. In every case the correction is a convention with a stated reference point, and the first question to ask of any of them is which reference point your laboratory used.