The standard therapeutic range for a phenytoin level is 10 to 20 micrograms per milliliter (μg/mL) of total drug in the blood, with the free (unbound) level ideally between 1 and 2 μg/mL. That range sounds simple, but phenytoin is one of the most difficult medications to dose correctly. Small changes in a patient’s albumin, kidney function, genetics, or co-medications can push blood levels from therapeutic to toxic or ineffective, sometimes without any change in dose. Understanding what a phenytoin level actually means requires knowing what the number captures and what it misses.
Why Phenytoin Levels Are Unusually Hard to Predict
Most drugs follow straightforward math: double the dose, roughly double the blood level. Phenytoin does not work this way. The liver enzymes responsible for breaking it down become saturated at the concentrations you typically see in clinical use, which means the drug shifts from predictable clearance to a pattern where even modest dose increases can cause disproportionate jumps in blood levels.1PubMed Central. Comparisons of Four Protein-Binding Models Characterizing the Pharmacokinetics of Unbound Phenytoin in Adult Patients Using Non-Linear Mixed-Effects Modeling This is why a patient stable at 12 μg/mL can suddenly climb to 25 μg/mL after what looks like a small bump in their daily dose.
This nonlinear behavior also extends to overdose situations. When someone’s level shoots well above the therapeutic range, the saturated enzymes cannot clear the drug efficiently, and the time it takes to return to normal stretches out dramatically. A person with a mildly elevated level might recover in a day or two, while someone with a severely toxic level could remain symptomatic for a week or more.2PubMed. Influence of bioavailability on the calculated Michaelis-Menten parameters of phenytoin in children This feature makes regular blood level monitoring essential in a way it simply is not for most other seizure medications.
Total Levels Versus Free Levels
When a lab reports a “phenytoin level,” it usually means the total concentration, which includes both the drug floating freely in blood and the roughly 90 percent that is bound to the protein albumin. Only the free (unbound) fraction is pharmacologically active; the bound portion is essentially parked and unavailable to cross into the brain. In a healthy person with normal albumin, the total level reliably reflects what the free level is doing. The trouble starts when albumin drops.
Patients in intensive care, those with liver disease, people who are malnourished, and anyone with significant kidney dysfunction often have low albumin. In these situations, more phenytoin is unbound and active, but the total level may look deceptively low. A total level of 8 μg/mL could correspond to a free level well within, or even above, the therapeutic range in a patient with an albumin of 2.5 g/dL. This mismatch is clinically dangerous: a clinician who only sees the low total might increase the dose, inadvertently pushing the patient into toxicity.
Studies comparing measured free levels to total levels in patients with low albumin consistently find that hypoalbuminemic patients show larger discrepancies between the two.3PubMed Central. Differences between the measured and calculated free serum phenytoin concentrations in epileptic patients The free phenytoin fraction is also significantly influenced by aging and kidney function, so older patients with even mild kidney impairment face a double risk of misleading total levels.4PubMed. Clinical evaluation of plasma free phenytoin measurement and factors influencing its protein binding
The Problem With Correction Equations
Because free phenytoin assays are not always available quickly, clinicians have long relied on mathematical correction equations to estimate the free level from the total level and the patient’s albumin. The most widely used is the Sheiner-Tozer (sometimes called Winter-Tozer) equation, which adjusts the total level based on albumin concentration. In theory, this gives you a “corrected” phenytoin level that should approximate what the total would be if albumin were normal.
In practice, the equation has serious limitations. A comprehensive review of studies evaluating its performance found the Sheiner-Tozer equation to be biased and imprecise across critical care, head trauma, and general neurology populations.5PubMed. A Comprehensive Review on the Predictive Performance of the Sheiner-Tozer and Derivative Equations for the Correction of Phenytoin Concentrations Another study looking specifically at critical care patients confirmed that both the original equation and its derivatives consistently overpredicted the true free phenytoin concentration, and that performance was especially poor in the sickest patients.6PubMed Central. Predictive Performance of the Winter-Tozer and Derivative Equations for Estimating Free Phenytoin Concentration
For patients with end-stage kidney disease on hemodialysis, the standard equation is even less reliable. One study concluded that the version of the equation designed for kidney disease was not accurate in predicting free phenytoin levels in dialysis patients, and a revised equation showed better accuracy but needs further validation.7PubMed. Characterization of Free Phenytoin Concentrations in End-Stage Renal Disease Using the Winter-Tozer Equation Patients with hypoalbuminemia and moderate or severe kidney dysfunction represent the hardest group to predict: in one analysis, none of the available correction methods achieved accuracy above 45 percent in patients with severe kidney dysfunction and low albumin.8PubMed. Predicting Unbound Phenytoin Concentrations: Effects of Albumin Concentration and Kidney Dysfunction
The practical takeaway is that whenever possible, a directly measured free phenytoin level is far more trustworthy than a calculated estimate, particularly in anyone with abnormal albumin, impaired kidneys, or critical illness. Correction equations remain a useful rough guide in routine outpatient settings with normal albumin, but they should not be the final word in complex patients.
How Genetics Shape the Dose You Need
Phenytoin is broken down in the liver primarily by two enzymes. About 90 percent of its metabolism runs through CYP2C9, with CYP2C19 handling the remaining 10 percent.9PubMed. Relative contributions of CYP2C9 and 2C19 to phenytoin 4-hydroxylation in vitro: inhibition by sulfaphenazole, omeprazole, and ticlopidine Genetic variations in these enzymes, which are common across populations worldwide, have a dramatic effect on how fast or slow a person clears phenytoin from their body.
People who carry reduced-function variants of CYP2C9 need substantially less phenytoin to reach the same blood levels. In one study, patients with at least one variant CYP2C9 allele required a mean dose roughly 37 percent lower than those with normal enzyme function to achieve therapeutic levels.10PubMed. The effect of genetic polymorphism of cytochrome P450 CYP2C9 on phenytoin dose requirement The difference was striking: nearly half of the variant carriers reached therapeutic concentrations on less than 200 mg per day, while most patients with normal enzymes needed over 300 mg per day.
When both CYP2C9 and CYP2C19 are impaired, the effect compounds. One pharmacokinetic study found that patients who were poor metabolizers for both enzymes had a maximum metabolic rate nearly 46 percent lower than normal metabolizers.11PubMed. Dosage recommendation of phenytoin for patients with epilepsy with different CYP2C9/CYP2C19 polymorphisms For these individuals, standard starting doses could push levels into the toxic range quickly.
Ancestry matters for which variants are most relevant. The CYP2C9*11 allele, found primarily in people of African descent, reduces enzymatic activity by more than 50 percent, resembling the effect of the well-known CYP2C9*3 “no-function” allele that is more common in European populations.12PubMed Central. The Impact of CYP2C9*11 Allelic Variant on the Pharmacokinetics of Phenytoin and (S)-Warfarin Certain CYP2C9 promoter variants linked to CYP2C19*2 can also affect how phenytoin induces its own metabolism over time, accounting for about 10 percent of the variation in maintenance dose across patients.13PubMed Central. CYP2C9*1B promoter polymorphisms, in linkage with CYP2C19*2, affect phenytoin autoinduction of clearance and maintenance dose Pharmacogenomic testing before starting phenytoin is increasingly advocated to reduce early toxicity risk, though it is still not routinely performed everywhere.
Drug Interactions and Tube Feeding Surprises
Phenytoin is both a victim and a perpetrator of drug interactions. It strongly induces several liver enzymes involved in drug metabolism, which means it can lower the blood levels of many other medications, including other seizure drugs like valproic acid, lamotrigine, and tiagabine.14The Lancet Neurology. Clinically significant antiepileptic drug interactions At the same time, many drugs can raise or lower phenytoin levels by competing for the same metabolic pathways or displacing it from albumin.
Valproic acid is a particularly tricky co-medication. It displaces phenytoin from its protein-binding sites, which raises the free phenytoin fraction. The total level may look unchanged or even drop, but the active free drug goes up. The result can be toxicity symptoms at a total level that appears to be safely within range.15PubMed. Pharmacologic interactions between valproate and other drugs This is one of the clearest examples of why checking a total phenytoin level alone can be misleading when certain interacting drugs are on board.
An often-overlooked interaction occurs in hospitalized patients receiving tube feeding. Enteral nutrition formulas administered through nasogastric tubes have been repeatedly associated with dramatically lower phenytoin levels. The mechanism involves both binding of the drug to the feeding tube itself at certain pH levels and reduced time for the drug to dissolve and be absorbed when feeding solutions move it quickly through the stomach.16PubMed. Phenytoin interaction with enteral feedings administered through nasogastric tubes This interaction is well documented in actual patients, though interestingly it has not been consistently reproduced in studies of healthy volunteers.17PubMed. Phenytoin and enteral feedings: does evidence support an interaction? The standard workaround in most hospitals is to hold tube feeding for one to two hours before and after giving phenytoin, though practice varies.
Recognizing Toxicity at Different Levels
Phenytoin toxicity does not happen all at once. Symptoms tend to appear in a rough progression that maps to how far above the therapeutic range the level climbs. The earliest sign is usually nystagmus, which is involuntary rhythmic eye movement, typically appearing as levels exceed 20 μg/mL. As levels rise further, ataxia (unsteady gait and coordination problems) sets in, generally around 30 μg/mL. Confusion, slurred speech, and lethargy follow at higher concentrations. One case report documented a patient with a level of 40 μg/mL presenting with confirmed ataxia.18PubMed. The Phenytoin Ataxia Enigma Unveiled “A Case Report” At extremely high levels, phenytoin can paradoxically cause seizures in someone taking it specifically to prevent them. A case of a child with a level of 80 μg/mL presented with generalized convulsions and reduced consciousness that initially mimicked a brain infection.19PubMed Central. Acute phenytoin intoxication in a 4-year-old mimicking viral meningoencephalitis
Cardiac toxicity from phenytoin is rare with oral dosing but is a real concern during intravenous administration, where rapid infusion can cause hypotension and abnormal heart rhythms. Deaths from phenytoin poisoning alone are uncommon. The bigger risk in severe overdose is prolonged hospitalization, since the saturated metabolism means the drug clears slowly, and a patient may remain confused or uncoordinated for days.
Phenytoin Levels in Pregnancy and Newborns
Pregnancy changes phenytoin handling in several ways that conspire to make level interpretation harder. Albumin concentrations fall as pregnancy progresses, which means more phenytoin is unbound. At the same time, the body’s total blood volume expands, and the rate of drug metabolism increases. The net effect is that both total and free phenytoin concentrations tend to drop during late pregnancy, even though the unbound fraction increases.20PubMed Central. Serum protein binding and free concentration of phenytoin and phenobarbitone in pregnancy For the pregnant person, this often means dose adjustments are needed in the third trimester, guided by free level measurements rather than total levels.
Phenytoin crosses the placenta freely. At the time of delivery, the drug is present in the newborn’s blood at concentrations similar to the mother’s.21PubMed. Anticonvulsants during pregnancy and lactation. Transplacental, maternal and neonatal pharmacokinetics Newborns, however, have lower total protein and albumin concentrations, which means their protein binding of phenytoin is reduced compared to older children and adults.22PubMed. Neonatal phenobarbital and phenytoin binding profiles A total phenytoin level measured in a newborn will underestimate the amount of active free drug. The same principle that applies to hypoalbuminemic adults applies even more acutely here.
Getting to Therapeutic Levels Quickly
When rapid seizure control is needed, phenytoin is given as an intravenous loading dose rather than started at a maintenance dose and allowed to accumulate. The goal is to push the blood level into the therapeutic range within an hour rather than waiting days. A study evaluating loading dose outcomes found that about 91 percent of patients achieved a corrected goal level after their first IV load, and that weight-based dosing was a strong predictor of success.23PubMed. Evaluation of Goal Phenytoin Levels After an Initial Intravenous Loading Dose at an Academic Medical Center Patients who reached their target had received a median dose of about 19 mg per kilogram, while those who fell short had received closer to 13 mg per kilogram.
Intravenous phenytoin comes with practical headaches. It must be given no faster than 50 mg per minute to limit the risk of low blood pressure and heart rhythm disturbances. It is highly alkaline and can cause severe vein irritation or tissue damage if it leaks outside the vein. It also tends to precipitate out of solution, so it requires an inline filter during infusion.24Advanced Emergency Nursing Journal. Optimizing Status Epilepticus Management in the Emergency Department: It’s About Time – Section: Phenytoin Fosphenytoin, a water-soluble prodrug that the body converts into phenytoin, was developed to avoid these infusion problems. It can be given faster and causes less vein irritation. However, a randomized comparison in routine emergency department use found that adverse events occurred at similar rates between the two formulations, and there was no significant difference in how long patients spent in the emergency department.25PubMed. Randomized evaluation of adverse events and length-of-stay with routine emergency department use of phenytoin or fosphenytoin
Bioavailability and Why the Same Dose Can Give Different Levels
Even when a patient takes the same milligram dose of phenytoin every day, the amount that actually reaches the bloodstream can vary. One early pharmacokinetic study found that the oral bioavailability of phenytoin capsules ranged from about 58 to 106 percent depending on how it was measured, with an average around 86 percent under steady-state conditions.26PubMed. Phenytoin: pharmacokinetics and bioavailability A comparative study of five different oral phenytoin formulations found bioavailability consistently between 68 and 74 percent, with no significant differences between the products.27Journal of Pharmacy and Pharmacology. A comparative study of the bioavailability of five different phenytoin preparations While the branded and generic oral products may be broadly interchangeable, the variability within any single formulation is wide enough that switching between products can still cause meaningful shifts in blood levels for an individual patient. This is why many guidelines recommend checking a phenytoin level any time a patient switches formulations.
Long-Term Side Effects That Show Up Gradually
Beyond acute toxicity, phenytoin taken over months or years can cause effects that do not correlate neatly with a single blood level. Gingival hyperplasia, an overgrowth of gum tissue, is one of the most well-known chronic side effects and is associated with phenytoin’s stimulation of collagen synthesis. The same mechanism may affect bone metabolism: phenytoin has been found to increase markers of bone formation in humans, which sounds beneficial but is linked to abnormal bone remodeling and, paradoxically, an increased risk of bone loss and fractures over time.28The Journal of Clinical Endocrinology & Metabolism. Phenytoin increases markers of osteogenesis for the human species in vitro and in vivo Long-term users are often advised to supplement calcium and vitamin D and to undergo periodic bone density screening.
Severe skin reactions, including Stevens-Johnson syndrome, are a rare but feared complication. Certain genetic markers, particularly the HLA-B*15:02 allele found more commonly in people of Southeast Asian ancestry, have been associated with an elevated risk of these reactions with another seizure drug, carbamazepine. For phenytoin, the relationship with HLA-B*15:02 is less clear-cut. A study in a Thai population found that the frequency of HLA-B*15:02 carriers among patients who developed severe skin reactions was not significantly different from controls.29Pharmacogenetics and Genomics. Associations between HLA class I and cytochrome P450 2C9 genetic polymorphisms and phenytoin-related severe cutaneous adverse reactions in a Thai population The FDA still recommends considering HLA-B*15:02 testing before starting phenytoin in at-risk populations, but the evidence linking this allele specifically to phenytoin skin reactions remains weaker than it is for carbamazepine.
When to Check a Level and What to Do With the Result
Routine phenytoin levels are typically drawn as trough measurements, meaning just before the next scheduled dose, to capture the lowest point in the dosing cycle. Common reasons to check a level include starting or changing the dose, adding or removing an interacting medication, a change in kidney or liver function, suspected non-adherence, new symptoms of toxicity, or breakthrough seizures despite a previously stable regimen. In critically ill patients, levels may be checked daily during the first few days of therapy.
When interpreting the result, the single most important question is whether a total level or a free level is needed for the clinical scenario. In outpatients with normal albumin and kidney function, a total level is usually sufficient. In anyone with low albumin, kidney disease, critical illness, pregnancy, or a co-medication that displaces phenytoin from proteins, a free level is more informative and should be specifically ordered. Relying on correction equations in these patients is a frequent source of dosing errors, and directly measuring the free level avoids that uncertainty.