What is the Winter-Tozer formula?

The Winter-Tozer formula adjusts a measured total phenytoin level for low albumin. Because phenytoin is heavily albumin-bound, hypoalbuminaemia makes the total level underestimate the active free drug. The formula is: corrected = measured divided by ((0.2 times albumin in g/dL divided by 4.4) plus 0.1).

When should I use the renal-failure variant?

Use the renal variant when creatinine clearance is below about 20 mL/min. Uraemia reduces phenytoin binding to albumin, so the adjustment factor drops from 0.2 to 0.1. Using the standard factor in severe renal failure would underestimate the free phenytoin and the risk of toxicity.

Why does low albumin matter for phenytoin?

Only free, unbound phenytoin is pharmacologically active. Roughly 90 percent of phenytoin is normally bound to albumin. When albumin falls, a larger fraction is free even though the measured total looks normal or low, so a patient can be toxic at a total level that appears subtherapeutic.

Is a corrected level as good as a free phenytoin assay?

No. A directly measured free (unbound) phenytoin level is the gold standard and should be used whenever it is available. The Winter-Tozer correction is an estimate for settings where only a total level and albumin are reported.

What albumin value is treated as normal?

This tool uses 4.4 g/dL as the normal reference, matching the original Winter-Tozer publication. Some centres use 4.0 g/dL; the difference shifts the corrected result only slightly. Enter albumin in g/L and the tool converts it to g/dL for you.

Phenytoin Dose Adjustment for Hypoalbuminaemia

Phenytoin level correction for low albumin

Phenytoin is about 90 percent bound to albumin, and only the free fraction is active. In sick, malnourished or critically ill patients with low albumin, the measured total phenytoin underestimates the active drug — a patient can be clinically toxic while the total level looks reassuringly low. This tool applies the Winter-Tozer formula to estimate the total level that would be seen if albumin were normal, with a separate variant for severe renal failure.

How it works

The standard correction is:

corrected = measured / ( (0.2 × albumin_g/dL / 4.4) + 0.1 )

The 0.2 term reflects normal albumin binding and 4.4 g/dL is the reference albumin. In end-stage renal failure (CrCl below ~20 mL/min), uraemic toxins displace phenytoin from albumin, so the binding term drops to 0.1:

corrected = measured / ( (0.1 × albumin_g/dL / 4.4) + 0.1 )

The tool accepts albumin in g/L (the common SI/UK unit) and divides by 10 to get g/dL before applying the formula.

Example and notes

A measured total phenytoin of 8 mg/L with an albumin of 25 g/L corrects to roughly 12 mg/L — moving a level that looked subtherapeutic into the therapeutic 10–20 mg/L band. That difference can change whether you increase the dose or hold it.

Where a laboratory can report a directly measured free phenytoin level, use that instead — it needs no correction. This calculator does not account for concurrent valproate, which also displaces phenytoin from albumin and would push the free fraction higher still.