Phenytoin TDM Calculator (Sheiner-Tozer Correction)

Medically reviewed by Dr. Taimoor Asghar, physician.

Correct total phenytoin levels for low albumin with the Sheiner-Tozer equation (including the end-stage renal disease variant), interpret the result against the 10 to 20 mg/L therapeutic range, and estimate loading doses and proportional maintenance dose corrections. Use the calculator above, then read the guide below to interpret your result and its limitations.

Why the measured phenytoin level can mislead

Phenytoin is one of the most protein-bound drugs in routine use: roughly 90 percent of the drug in blood is attached to albumin, leaving only about 10 percent free to cross the blood-brain barrier and exert an effect. Laboratories report the total level, which is the sum of bound and free drug. In a healthy person with normal albumin this is fine, because the free fraction stays near 10 percent and the familiar therapeutic range of 10 to 20 mg/L (equivalent to 10 to 20 micrograms/mL) was defined against total levels in such patients.

The problem starts when albumin falls, which is common in critically ill patients, people with liver disease, nephrotic syndrome, malnutrition, burns, pregnancy, and older adults. With less albumin available, a larger share of each phenytoin molecule stays free. The total level then drops even though the free, active concentration is unchanged or rising. A measured total of 8 mg/L in a patient with an albumin of 2.5 g/dL looks subtherapeutic on the lab slip, but the corrected level is about 13.3 mg/L, squarely therapeutic. Increasing the dose on the basis of the uncorrected number risks pushing the free level into toxicity.

This trap is especially dangerous because the clinical picture can be confusing. A patient with a low total level and ongoing seizures seems to need more drug, yet the seizures may be breaking through for reasons unrelated to phenytoin, while the free level is already adequate. Conversely, a patient with a low total level who is drowsy, ataxic, or has nystagmus is showing you the free level directly: trust the patient, not the uncorrected number. Whenever albumin is abnormal, the total level should never be interpreted without correction or a free level.

The Sheiner-Tozer equation

Sheiner and Tozer published their correction in 1978 after studying phenytoin binding across a range of albumin concentrations. The equation is: corrected phenytoin equals the measured total level divided by (0.2 times the albumin in g/dL, plus 0.1). The denominator estimates the free fraction relative to normal: at an albumin of 4.4 g/dL the denominator is 0.98, so the correction barely changes the number, while at an albumin of 2.5 g/dL the denominator is 0.60 and the corrected level is two-thirds higher than measured.

A worked example makes the arithmetic concrete. A patient has a measured total phenytoin of 8 mg/L and an albumin of 2.5 g/dL. The denominator is (0.2 times 2.5) plus 0.1, which is 0.6. Dividing 8 by 0.6 gives 13.3 mg/L. The uncorrected value suggests the dose should rise; the corrected value says the patient is therapeutic and the dose should stay where it is. This calculator performs exactly this arithmetic and then places the corrected value into its clinical band.

The equation has limits worth knowing. It was derived in patients without severe renal failure, it assumes binding is otherwise normal, and it becomes less reliable at very low albumin levels or when other highly bound drugs compete for binding sites. Valproic acid, for instance, displaces phenytoin from albumin and also inhibits its metabolism, which the equation does not capture. Treat the corrected level as a much better estimate than the raw total, not as a perfect measurement of free drug.

The end-stage renal disease variant

In end-stage renal disease, uraemic toxins compete with phenytoin for albumin binding sites and also alter the binding affinity itself, so the free fraction rises even more than the albumin alone would predict. For these patients a modified equation is used: corrected phenytoin equals the measured level divided by (0.1 times the albumin in g/dL, plus 0.1). At an albumin of 3.0 g/dL the denominator is 0.40, meaning the corrected level is two and a half times the measured one.

Take a dialysis patient with a measured level of 6 mg/L and an albumin of 3.0 g/dL. The standard equation would give 6 divided by 0.7, about 8.6 mg/L, still subtherapeutic. The ESRD variant gives 6 divided by 0.4, which is 15 mg/L, comfortably therapeutic. Using the wrong equation here would lead to a dose increase in a patient who needs none. Select the end-stage renal disease option in the calculator whenever the patient has an estimated glomerular filtration rate below about 15 mL/min or is on dialysis.

Even the ESRD variant is an approximation, and most toxicologists prefer a directly measured free phenytoin level in dialysis patients when the assay is available. The therapeutic range for free phenytoin is 1 to 2 mg/L. If the free level can be obtained, it supersedes any correction equation. Use the equation when the free assay is unavailable or while waiting for its result.

Therapeutic range and toxicity bands

The accepted therapeutic range for total (or corrected) phenytoin is 10 to 20 mg/L. Within this range most patients achieve seizure control without toxicity, though individual patients vary and some need levels slightly outside it. The range is a population guide, not a personal prescription: a patient who is seizure-free at 9 mg/L should not have the dose pushed up to chase the number, and a patient seizing at 18 mg/L may need a different drug rather than a higher level.

Above 20 mg/L the classic dose-related signs appear in a predictable sequence. Nystagmus, particularly on lateral gaze, is usually the first sign and can appear just above 20 mg/L. Ataxia and slurred speech follow in the 30 mg/L region. Beyond 40 mg/L consciousness is depressed, and very high levels can cause seizures paradoxically, coma, and cardiovascular collapse, especially with rapid intravenous loading. This calculator bands the corrected level accordingly so the number arrives with its clinical meaning attached.

Chronic toxicity looks different from acute overdose. Gingival hyperplasia, hirsutism, coarsened facial features, peripheral neuropathy, cerebellar atrophy, megaloblastic anaemia from folate depletion, and reduced bone mineral density develop over months to years of therapy. These do not correlate tightly with any single level, which is why long-term monitoring includes clinical review and not just numbers. Any patient on chronic phenytoin should have folate, vitamin D, and bone health considered as part of routine care.

Loading doses and the volume of distribution

When seizures need rapid control, a loading dose brings the level into the therapeutic range quickly instead of waiting the five to seven days that maintenance dosing alone would take. The loading dose depends on how far the current level is from target and on the drug's volume of distribution, which for phenytoin is about 0.65 L/kg. The formula is: loading dose in mg equals (target level minus current level) times 0.65 times body weight in kg.

For a 70 kg patient with a level of 5 mg/L and a target of 15 mg/L, the gap is 10 mg/L, and the loading dose is 10 times 0.65 times 70, which is 455 mg. In practice this is rounded to a convenient amount such as 500 mg and given intravenously, slowly, with cardiac and blood pressure monitoring, because rapid phenytoin infusion can cause hypotension and arrhythmia. Fosphenytoin, the phosphate prodrug, is often preferred for loading because it can be given faster and intramuscularly, with doses expressed in phenytoin equivalents.

Oral loading is possible but slower and less predictable, with peak levels several hours after the dose. Whatever the route, a level should be checked after loading to confirm the target was reached before maintenance dosing continues. Do not use the loading estimate when the measured level already exceeds the target: the calculator returns no loading dose in that situation, because adding drug to a therapeutic or toxic level is never indicated.

Adjusting the maintenance dose

The simplest maintenance adjustment is proportional: new daily dose equals current daily dose times (target level divided by measured level). A patient on 300 mg daily with a corrected level of 10 mg/L and a target of 15 mg/L would move to about 450 mg daily. This calculator shows that estimate when you enter the current dose. It is a reasonable starting point, but phenytoin's non-linear kinetics demand caution with it.

Phenytoin is cleared by the liver enzymes CYP2C9 and CYP2C19, which begin to saturate within the therapeutic range. This Michaelis-Menten behaviour means the relationship between dose and level is not a straight line: near the top of the range, a small dose increase produces a disproportionately large level increase. The classic teaching example is that raising the dose from 300 to 400 mg daily can take a level from 12 to well above 20 mg/L in a susceptible patient. Dose changes should therefore be modest, typically 25 to 50 mg steps in adults, with a recheck before further adjustment.

Steady state matters for timing. After any dose change, about five to seven days (roughly five half-lives, with phenytoin's half-life averaging around 24 hours but varying widely) are needed before the new level reflects the new dose. Checking too early captures a level still in transition and invites overcorrection. Trough levels, drawn just before the next dose, are the most consistent for comparison over time, and levels should always be compared like with like.

When to measure free phenytoin instead

The correction equations handle the common case of low albumin well, but several situations call for a directly measured free level. End-stage renal disease is the classic one, as discussed above. Others include pregnancy, where albumin falls and clearance rises; critical illness with multi-organ dysfunction; hypoalbuminaemia combined with interacting drugs such as valproic acid; and any patient whose clinical picture contradicts the corrected number, for example toxicity signs with a therapeutic corrected level, or uncontrolled seizures with a high corrected level.

The therapeutic range for free phenytoin is 1 to 2 mg/L, reflecting the roughly 10 percent free fraction at normal binding. A free level inside this range generally means dosing is appropriate regardless of what the total or corrected numbers say. Note that some laboratories report free phenytoin in different units, so check whether your lab uses mg/L or micrograms/mL before comparing: they are numerically identical (1 mg/L equals 1 microgram/mL), but mixing units across reports is a classic source of error.

Free levels are also the right tool when binding is altered by something the equations cannot see. High bilirubin, uraemic toxins beyond what the ESRD equation models, and displacement by other highly bound drugs all change the free fraction unpredictably. In these cases the equations give a false sense of precision. When in doubt, and especially before making a large dose change in a complex patient, the measured free level is the tiebreaker.

Sources

• Sheiner LB, Tozer TN. Clinical pharmacokinetics: concepts and applications (phenytoin protein binding correction). Clinical Pharmacokinetics. 1978;3(4):263-280. doi:10.2165/00003088-197803040-00002
• Patsalos PN, et al. Antiepileptic drugs: best practice guidelines for therapeutic drug monitoring. Epilepsia. 2008;49(7):1239-1276. doi:10.1111/j.1528-1167.2008.01561.x

Frequently asked questions

What is a normal phenytoin level?

The accepted therapeutic range for total phenytoin is 10 to 20 mg/L, which is numerically identical to 10 to 20 micrograms/mL. When albumin is low, interpret the Sheiner-Tozer corrected level against this range, not the raw measured total. For directly measured free phenytoin, the therapeutic range is 1 to 2 mg/L.

Why does low albumin change the phenytoin result?

About 90 percent of phenytoin in blood is bound to albumin and inactive; only the free 10 percent acts on the brain. When albumin falls, more drug stays free, so the total level drops while the active concentration stays the same or rises. The lab reports the total, which then underestimates the drug effect. Correcting with Sheiner-Tozer, or measuring the free level directly, reveals the true position.

What is the Sheiner-Tozer equation?

It corrects a total phenytoin level for hypoalbuminaemia: corrected level equals measured level divided by (0.2 times serum albumin in g/dL, plus 0.1). For example, a measured level of 8 mg/L with an albumin of 2.5 g/dL corrects to 13.3 mg/L. In end-stage renal disease a variant denominator, (0.1 times albumin plus 0.1), is used instead.

When should I use the ESRD version of the correction?

Use the end-stage renal disease variant when the patient has an estimated glomerular filtration rate below about 15 mL/min or is on dialysis. Uraemic toxins displace phenytoin from albumin beyond what the albumin value alone predicts, so the standard equation under-corrects. Even better in these patients is a directly measured free phenytoin level (therapeutic 1 to 2 mg/L).

Can I adjust the phenytoin dose proportionally to the level?

As a first estimate, yes: new dose equals current dose times (target level divided by current level). But phenytoin clearance saturates within the therapeutic range, so near the top of the range a small dose increase can cause a large level jump. Make changes in modest steps, recheck a trough level after 5 to 7 days at the new dose, and always interpret corrected or free levels, not raw totals, when albumin is abnormal.

What are the signs of phenytoin toxicity?

Acute dose-related toxicity follows a sequence: nystagmus on lateral gaze just above 20 mg/L, then ataxia and slurred speech around 30 mg/L, then drowsiness and depressed consciousness above 40 mg/L. Rapid intravenous loading can also cause hypotension and cardiac arrhythmia, which is why loading doses are given slowly with monitoring. Chronic toxicity includes gum overgrowth, excess hair growth, peripheral neuropathy, and reduced bone density.

Clinical sources: Sheiner LB, Tozer TN. Clinical Pharmacokinetics. 1978;3(4):263-280; Patsalos PN et al. Epilepsia. 2008;49(7):1239-1276.