Phenytoin Dosing in Hypoalbuminaemia, Renal Impairment and Valproate Co-therapy
A 54-year-old man with focal epilepsy, seizure-free for three years on phenytoin 300 mg once daily, is admitted with community-acquired pneumonia and acute-on-chronic kidney disease. Serum creatinine is 2.9 mg/dL (baseline 1.3), eGFR 25 mL/min/1.73 m², albumin 2.0 g/dL, bilirubin normal, INR 1.1. For six weeks he has also taken valproate 500 mg twice daily, added by a neurologist for a suspected second seizure type. Trough total phenytoin is 7.5 mg/L (reference 10–20 mg/L). The admitting resident plans to increase phenytoin to 400 mg daily. A free phenytoin level, drawn before any change, returns 2.4 mg/L (reference 1.0–2.0 mg/L). The patient is drowsy but has no nystagmus, and the drowsiness is being attributed to pneumonia.
1. Apply the Sheiner–Tozer correction, interpret the result, and state the limits of the equation in this patient.
Read the worked answer
Corrected total phenytoin = measured total ÷ (0.2 × albumin + 0.1) = 7.5 ÷ (0.2 × 2.0 + 0.1) = 7.5 ÷ 0.5 = 15.0 mg/L
So the total concentration, which appears grossly subtherapeutic, corresponds to a corrected total of 15 mg/L - mid-therapeutic. This agrees with the measured free level of 2.4 mg/L, which is above the 1.0–2.0 mg/L reference range and therefore suggests a mildly toxic free concentration, consistent with the drowsiness that is being wrongly attributed to pneumonia.
Limits of the equation: the 0.1 term assumes normal renal function. In advanced renal impairment, accumulated uremic toxins and organic acids displace phenytoin further, and the equation underestimates the free fraction. Albumin quality is also altered in acute illness. In this patient the directly measured free level is the more reliable number, and the total level is a poor surrogate.
2. Decide on the proposed dose increase and specify the monitoring plan, including what to do about attribution of the drowsiness.
Read the worked answer
Do not increase the dose. 400 mg would raise exposure by roughly one third, pushing an already supra-reference free level into the frankly toxic range and increasing the risk of nystagmus, ataxia and confusion, with no seizure-control benefit because the free concentration is already above the therapeutic window.
Plan:
- Continue phenytoin 300 mg daily and record a free phenytoin level daily during the acute illness, with clinical toxicity checks each shift.
- Re-examine the patient specifically for nystagmus and ataxia at the bedside; if present, the drowsiness is drug-related, not pneumonic.
- Recheck the free level after renal function recovers, because the free fraction will fall as uremic displacers clear and the same dose may then become subtherapeutic.
- Document in the record that the total-level reference range does not apply to this patient, so the next clinician does not repeat the error.
- Review sedating co-medications before attributing drowsiness to infection.
3. Explain valproate's contribution to these numbers and what happens to free and total phenytoin if valproate is stopped abruptly.
Read the worked answer
Valproate acts in two ways: it inhibits phenytoin metabolism (contributing to the raised free concentration) and, being highly protein-bound, it competes with phenytoin for albumin binding sites. With albumin already at 2.0 g/dL, this competition shifts a larger fraction of phenytoin into the free, pharmacologically active pool. Free phenytoin is what equilibrates with brain, so the free level is the parameter that matters clinically.
If valproate is stopped abruptly: the free fraction falls, free phenytoin drops even at an unchanged dose, and there is a seizure risk; abrupt withdrawal of an antiepileptic in a patient with epilepsy is itself hazardous. Total phenytoin may appear to rise as binding sites free up, which can mislead if only total levels are followed.
Therefore: do not stop valproate unilaterally. Involve neurology, and if a change is needed, make it deliberately and recheck free phenytoin about 5–7 days later, once the interaction has washed out. Titrate against free levels, not total.