Corrected Phenytoin Calculator

Use the Corrected Phenytoin Calculator

Free corrected phenytoin calculator using the Sheiner-Tozer equation to adjust a low phenytoin level for albumin or renal failure and estimate free drug.

How to Use Corrected Phenytoin Calculator

1. Step 1: Choose your unit system

   Select US (mg/L for phenytoin, g/dL for albumin) or SI (micromol/L and g/L). Note that mg/L and mcg/mL are the same number, so no conversion is needed between those two.

2. Step 2: Enter the measured total phenytoin

   Type the trough total phenytoin concentration from the lab report into the Measured Total Phenytoin field. Do not enter a free or unbound phenytoin result here, because that value is already corrected.

3. Step 3: Enter the serum albumin

   Enter the albumin drawn at or near the same time as the phenytoin level in the Serum Albumin field. An albumin more than 48 hours old can shift the corrected result by 15 to 20 percent.

4. Step 4: Select the renal function and equation variant

   Pick Standard for a creatinine clearance of 20 mL/min or higher, Renal failure for a clearance under 10 mL/min or dialysis, or the revised 0.25 coefficient for neurocritical care patients.

5. Step 5: Read the corrected level and reclassification alert

   The corrected phenytoin level appears instantly with a therapeutic band badge. The alert below it tells you whether the correction moved the result into a different band than the reported value.

6. Step 6: Compare the equations and check the free estimate

   Review the comparison table to see how all three published equations score the same inputs, then check the estimated free phenytoin against the 1 to 2 mg/L target before any dose decision.

Key Features

• Sheiner-Tozer correction with standard (0.2), revised (0.25), and renal-failure (0.1) coefficients
• Side-by-side table showing how all three published equations score the same inputs
• Estimated free phenytoin concentration and unbound fraction against the 1-2 mg/L target
• Alert when the correction moves the result into a different therapeutic band
• Dual-marker range bar plotting the measured and corrected levels together
• US (mg/L, g/dL) and SI (micromol/L, g/L) unit support

Understanding Results

Formula

The Sheiner-Tozer equation divides the measured total phenytoin by a correction factor built from serum albumin. The factor estimates how much of a healthy adult's albumin binding capacity remains, so a factor of 0.52 means only 52% of normal binding is available and the reported total understates true drug exposure by roughly half.

Albumin is entered in g/dL and phenytoin in mg/L (identical to mcg/mL). The coefficients are deliberately scaled: 0.2 × 4.5 + 0.1 equals exactly 1.0, so a patient with a normal albumin of 4.5 g/dL receives no adjustment. The trailing 0.1 represents residual binding to plasma proteins other than albumin and keeps the denominator from collapsing toward zero at extreme hypoalbuminemia.

Reference Ranges & Interpretation

The adult therapeutic range for total phenytoin is 10–20 mg/L (40–79 µmol/L), and for free phenytoin 1–2 mg/L. Compare the corrected value against those bands, not the reported one. Concentration-related effects follow a predictable sequence: horizontal nystagmus near 20 mg/L, ataxia and slurred speech above 30 mg/L, and lethargy or encephalopathy above 40 mg/L.

The single most useful output is the reclassification alert. When a reported 6.5 mg/L at an albumin of 2.1 g/dL corrects to 12.5 mg/L, the result crosses from subtherapeutic into therapeutic, and the appropriate action flips from loading the patient to leaving the dose alone. Because phenytoin follows saturable Michaelis-Menten kinetics inside the therapeutic range, a modest dose increase applied to a misread level can produce a disproportionately large rise in serum concentration.

Assumptions & Limitations

The equation assumes a steady-state trough total phenytoin drawn with a same-day albumin, normal albumin binding affinity, and no competing highly protein-bound drugs. Allow 5–7 days after any dose change before treating a level as steady state.

It is not valid with concurrent valproate, which displaces phenytoin from albumin and raises the free fraction to 20–30% while albumin still reads normal. It has not been validated in pregnancy, in neonates, or at albumin values below about 1.5 g/dL, where the constant term dominates. A directly measured free phenytoin concentration by ultrafiltration outperforms this estimate and should be ordered whenever renal failure is present, an interacting drug is on board, the corrected value sits close to a decision threshold, or the clinical picture and the number disagree. Never adjust an antiepileptic regimen without the prescribing clinician.

Complete Guide: Corrected Phenytoin Calculator

A corrected phenytoin calculator exists because roughly 90% of phenytoin in the bloodstream is bound to albumin, and only the remaining 10% crosses into the brain to stop seizures. Drop albumin from 4.5 g/dL to 2.1 g/dL and the total level a lab reports can fall by nearly half while the pharmacologically active free drug has not moved at all. That is the trap: a patient in the intensive care unit with a reported level of 6.5 mg/L looks badly subtherapeutic, gets a loading dose, and ends up genuinely toxic within 48 hours.

The Sheiner-Tozer equation — usually called Winter-Tozer in pharmacy practice — converts that misleading total into the level you would have seen if albumin were normal. What most references gloss over is that there is not one equation. There are at least three in active clinical use, they disagree with each other by 40% or more at low albumin, and picking the wrong one is its own dosing error. This guide compares them directly.

Three equations, three different answers

Every version shares the same skeleton: divide the measured total level by a correction factor built from albumin. Only the binding coefficient changes, and that single number carries all the disagreement.

Run one patient through all three and the divergence is stark. At an albumin of 2.1 g/dL and a reported level of 6.5 mg/L, the standard equation returns 12.5 mg/L, the revised coefficient returns 10.4 mg/L, and the renal variant returns 21.0 mg/L. Those three numbers sit in three different clinical bands: comfortably therapeutic, barely therapeutic, and above range. The calculator above prints all three side by side for exactly this reason — the spread is information, not noise.

What the denominator is actually doing

The correction factor is an estimate of the fraction of normal binding capacity still available. Read it as a percentage. A factor of 0.52 means this patient retains 52% of the albumin binding a healthy adult would have, so the measured total under-represents the true drug exposure by roughly half.

The trailing + 0.1 is not decoration. It is the floor: even with albumin at zero, a residual 10% of binding is assumed to come from other plasma proteins, which keeps the equation from dividing by something near zero and producing an absurd number. That floor is also why the correction misbehaves below about 1.5 g/dL — the constant starts to dominate a shrinking albumin term, and the output becomes more arithmetic than physiology.

Notice the coefficients are scaled deliberately. Plug 4.5 g/dL into the standard equation: 0.2 × 4.5 + 0.1 = 1.0 exactly, so a patient with normal albumin gets no adjustment. The revised coefficient hits 1.0 at 3.6 g/dL instead, which is why it always returns a lower corrected value than the standard version for the same inputs.

Worked example: albumin 2.1 g/dL, level 6.5 mg/L

A 68-year-old admitted with aspiration pneumonia has been on phenytoin 300 mg daily for years. Trough total phenytoin comes back at 6.5 mg/L against a target of 10–20 mg/L. Albumin is 2.1 g/dL. Serum creatinine is normal.

1. Pick the equation. Renal function is intact, so the standard coefficient of 0.2 applies.
2. Build the correction factor. 0.2 × 2.1 = 0.42, plus 0.1 gives 0.52.
3. Divide. 6.5 ÷ 0.52 = 12.5 mg/L corrected.
4. Estimate the free level. 0.1 × 12.5 = 1.25 mg/L, sitting inside the 1–2 mg/L free target.
5. Derive the free fraction. 1.25 ÷ 6.5 = 19.2% unbound, roughly double the normal 10%.

The reported 6.5 mg/L reads as subtherapeutic. The corrected 12.5 mg/L reads as therapeutic. That reclassification is the entire clinical payload: this patient does not need more phenytoin, and giving a 500 mg load here would push the free level toward 2 mg/L and beyond. Because phenytoin follows saturable Michaelis-Menten kinetics near therapeutic concentrations, a 30% dose increase in this range can produce a far larger jump in serum level than a proportional change would predict — there is very little forgiveness on the way up.

When to switch to the renal-failure variant

Uremia does something albumin concentration alone cannot capture. Retained organic acids compete for the same binding sites on albumin and structurally alter the protein, so a dialysis patient with a perfectly respectable albumin of 4.0 g/dL still runs a free fraction closer to 20–25% rather than 10%. The renal variant handles this by halving the coefficient to 0.1, which produces a 1.8–2.0× adjustment even at normal albumin.

The usual threshold is a creatinine clearance under 10 mL/min or any patient on hemodialysis, though many institutions apply it below 20 mL/min. If you are unsure which side of that line a patient falls on, calculate it rather than eyeball it — the creatinine clearance calculator uses the Cockcroft-Gault equation that these thresholds were originally defined against, while an eGFR calculator gives the body-surface-normalized figure used for CKD staging. They are not interchangeable at the extremes of body weight, and the phenytoin literature is written in creatinine clearance.

Corrected phenytoin calculator vs. a measured free level

A directly measured free phenytoin concentration beats any equation, and it is not close. Ultrafiltration assays measure the unbound fraction rather than inferring it, which matters because the correction carries a wide error margin: published comparisons find the Winter-Tozer estimate falls outside a clinically acceptable window of the measured free level in a substantial minority of critically ill patients. The equation is a screening tool, not a substitute.

So why is the correction still used daily? Turnaround and availability. Free phenytoin assays are sent out at many hospitals with a 24–72 hour result window, while total phenytoin and albumin come back the same morning. When a decision has to happen before rounds end, the corrected value is what exists. The practical rule most pharmacists apply: use the correction to decide whether the situation is ambiguous enough to justify ordering a free level, and order the free level whenever the corrected value sits within about 20% of a decision threshold, whenever there is renal failure, or whenever the clinical picture and the number disagree.

Valproate: the displacement the equation cannot fix

Valproic acid binds the same albumin sites as phenytoin and wins. Co-administration can push the phenytoin free fraction to 20–30% while albumin sits at a completely normal 4.2 g/dL — and because the equation only reads albumin concentration, it will report a correction factor of about 0.94 and tell you nothing is wrong. The displacement is invisible to the arithmetic.

The effect is also time-dependent, which makes it worse. Acute displacement raises free phenytoin sharply; over the following days, increased clearance of the newly unbound drug pulls the total level down while free concentrations settle somewhere above baseline. A total level checked a week into combination therapy can look low and stable while the free level is the one causing the patient's nystagmus. Any patient on both drugs needs a measured free level, full stop. The same logic applies to other highly protein-bound competitors, including warfarin, salicylates at anti-inflammatory doses, and ceftriaxone. When multiple interacting drugs are in play, a drug half-life calculator helps map how long each competitor stays at concentrations high enough to matter.

Four mistakes that break the correction

Correcting a free level. The equation takes a total phenytoin concentration as input. Feeding it a free level that is already unbound double-corrects and inflates the result by roughly tenfold. If the lab report says "phenytoin, free" or the value is under 3 mg/L in a treated adult, stop.

Using a stale albumin. Albumin has a half-life near 20 days in health but falls fast in acute illness through capillary leak and dilution. An albumin drawn five days before the phenytoin level can be 0.8 g/dL off, which at low albumin shifts the corrected result by 15–20%. Draw them together.

Correcting a non-trough sample. Phenytoin has a variable half-life of 12–36 hours that lengthens as concentrations rise. A level drawn two hours after an oral dose reflects absorption, not steady state, and no albumin correction rescues a badly timed sample. Wait for a true trough, and allow 5–7 days after a dose change before treating any level as steady state.

Applying it to fosphenytoin without conversion. Fosphenytoin is dosed in phenytoin equivalents and must be fully converted to phenytoin before a meaningful level exists, which takes about 15 minutes after an intravenous dose and longer intramuscularly. Sampling too early measures a prodrug that has not become the drug yet. The same discipline that applies to any weight-based regimen in a dosage calculator applies here: confirm which molecule the number describes before doing arithmetic on it.

One structural note worth carrying beyond phenytoin. This is the same reasoning a corrected calcium calculator applies to serum calcium, but the arithmetic runs in the opposite direction. Calcium correction adds a fixed amount per gram of missing albumin; phenytoin correction divides by a proportional binding factor. Substances that bind albumin proportionally scale; those measured as a pool with a bound compartment shift additively. Confusing the two produces answers that are not merely imprecise but wrong by a large factor.

References

1. National Library of Medicine, StatPearls. Phenytoin: pharmacology, therapeutic monitoring, and toxicity.
2. U.S. Food and Drug Administration. Dilantin (phenytoin sodium) prescribing information — protein binding, therapeutic range, and saturable kinetics.
3. National Institute of Neurological Disorders and Stroke. Epilepsy and seizures information page.

Frequently Asked Questions