Nystagmus and ataxia on phenytoin: interpreting a level in the presence of hypoalbuminaemia
A 26-year-old man, 62 kg, has been taking phenytoin 300 mg at bedtime for six weeks after a post-traumatic seizure. Two weeks ago a locum doctor added 100 mg because of a single breakthrough seizure, and the man admits he sometimes forgets a dose and doubles the next one. He now presents with horizontal nystagmus, ataxia, dysarthria and drowsiness. Investigations: total phenytoin 34 micrograms/mL (therapeutic range 10-20), serum albumin 2.9 g/dL, ALT 32 U/L, creatinine 1.0 mg/dL, haemoglobin 13.2 g/dL. The sample was drawn at 08:00, twelve hours after the last dose. He is on no other medicines, drinks no alcohol, and his seizures have been controlled since the dose change. He lives four hours from the hospital and asks whether he can simply halve his tablets for a few days and go home.
1. Interpret the phenytoin level. Is the albumin correction necessary here, and does the result explain the symptoms?
Read the worked answer
The measured total concentration of 34 micrograms/mL is already above the upper therapeutic limit of 20, and the albumin correction makes the exposure far worse.
Phenytoin is about 90% bound to albumin, and it is the free fraction that is pharmacologically active. With hypoalbuminaemia the total level under-reads the free concentration. Using the Sheiner-Tozer correction for a patient with normal renal function:
corrected concentration = measured / (0.2 x albumin + 0.1) = 34 / (0.2 x 2.9 + 0.1) = 34 / 0.68 = 50 micrograms/mL
That is 2.5 to 5 times the upper therapeutic limit, and it matches the clinical picture exactly: nystagmus appears around 20, ataxia and dysarthria in the high 20s to 30s, and drowsiness and confusion as levels climb further. The timing is right (sample drawn at 12 hours, close to trough, so the peak was higher still), there is no alternative cause (normal ALT, normal creatinine, no alcohol, no other drugs), and the reaction is dose-related and reversible. This is phenytoin toxicity, not an idiosyncratic reaction.
Two things must be said about the correction, though. It is a tool for interpretation, not a target: at very high total concentrations the correction over-estimates free drug, and the decision to treat should be driven by the patient's clinical state. If free phenytoin measurement is available, it would settle the question. And the underlying diagnosis - hypoalbuminaemia at 2.9 g/dL in a young man with a head injury on phenytoin deserves its own explanation (nutrition, inflammation, protein loss).
2. Explain why a 100 mg dose increase produced this degree of toxicity, and set out the immediate and restart plan.
Read the worked answer
Why 33% became several-fold. Phenytoin metabolism follows Michaelis-Menten (saturable) kinetics. At therapeutic doses many patients are already operating near the enzyme's Vmax, so the relationship between dose and steady-state concentration is not linear: a 33% increase in dose (300 to 400 mg) can roughly double or triple the steady-state level. Combine that with (a) missed-then-doubled doses, which create peaks well above the steady state, and (b) an albumin of 2.9 g/dL, which raises the free fraction, and a modest dose change becomes a major exposure change. This is why phenytoin dose increments should be 25-30 mg at a time, with a level 5-7 days later - longer when levels are high, because half-life lengthens as the enzymes saturate.
Immediate management: admit; stop phenytoin; supportive care with airway protection, continuous monitoring and IV fluids as needed; check glucose and electrolytes; there is no antidote, and phenytoin is not usefully dialysable. Symptoms resolve over several days as the level falls, slowly, because clearance is saturated. Do not send him home to halve his tablets: he is ataxic after a head injury, he lives four hours away, he has already demonstrated that he mismanages doses, and the decline in level needs to be observed.
Restart plan: once nystagmus and ataxia have resolved and the level has fallen to about 15 micrograms/mL, restart at roughly 4 mg/kg/day (about 250 mg/day) - that is, below the dose that produced toxicity - and recheck a trough level at 5-7 days with albumin correction. If levels prove unstable on twice-daily or once-daily dosing, consider switching to an agent with linear kinetics (levetiracetam or carbamazepine), cross-tapering carefully and counselling about adherence.
3. What should be reported to the pharmacovigilance system, and how would you argue the case for reporting a 'known' dose-related effect?
Read the worked answer
What to report. A suspected adverse drug reaction report containing: patient identifiers and age/weight; the drug, dose (400 mg at night), indication, start dates of the original dose and the increment; concomitant medicines (none) and relevant comorbidity (head injury, hypoalbuminaemia); the reaction - nystagmus, ataxia, dysarthria, drowsiness - with date of onset and severity; the total phenytoin 34 micrograms/mL with albumin 2.9 g/dL, the albumin-corrected value of about 50, and the timing of the sample; the action taken (drug stopped) and the outcome; and the reporter's details.
Seriousness: the reaction caused admission to hospital, so it meets the seriousness criterion, regardless of whether it is 'expected'.
Causality, argued with a standard scheme (WHO-UMC): plausible time to onset after the dose increment; a recognised concentration-related effect of phenytoin; improvement on withdrawal; no alternative explanation in the history, examination or laboratory data; and confirmation by a measured level. That places it at 'certain' or 'probable' rather than 'possible'.
Why report a known effect:
- It is a preventable medication error as much as an ADR - a dose escalation without level monitoring, in a patient with a low albumin, plus erratic adherence. Reporting captures the system failure, not just the pharmacology.
- Known reactions are exactly where preventable harm concentrates; the value of the report is in showing that the risk was not managed, not in discovering a new effect.
- Aggregated reports document the magnitude of the dose-concentration problem locally and can support a protocol change - for example, mandatory albumin correction and a 5-7 day level after any phenytoin increment, introduced through the hospital's pharmacy and therapeutics committee.
Route it through the hospital's ADR unit to the national pharmacovigilance programme coordinated by the DDA, and record in the patient's notes that he has had phenytoin toxicity.