All antiepileptic drug blood level determinations at the University of Virginia Medical Center were analyzed over a 13 month period. Those values with unbound fractions outside the expected range and discordance between unbound and total levels were compared to seizure control and occurrence of side effects. Unbound phenytoin (PHT) levels were clinically significant (outside the expected range, a discordant ratio of unbound to total level, and seizures or side effects necessitating dosage change) in 61 of 254 determinations (24%) and 44 of 155 patients (28%). The clinical significance involved current seizures for 52 values and side effects for 24 (both for 15). The frequency of clinically significant values was similar with PHT monotherapy (20%) and polytherapy (27%). The rates of current seizures and side effects were also similar in these 2 subgroups. For valproate (VPA), 15% of unbound values from 18% of patients were clinically significant with regard to current seizures and side effects and the rate was similar for monotherapy and polytherapy. There were no clinically significant unbound levels for carbamazepine. 
COMMENT. These data suggest that some patients receiving phenytoin or valproate have unbound drug levels and ratios that may add information of importance to improvement of seizure control and reduction of side effects. The authors suggest that patients receiving PHT as mono- or polytherapy be monitored with either unbound blood level measurements only or have unbound blood levels regardless of the total level if there is an unresolved clinical problem. The expected unbound fraction of PHT is 8-12%.
In a 17-year-old girl with thrombotic thrombocytopenia purpura who was treated by plasmapheresis the free and total concentrations of phenytoin used for the treatment of a coincident seizure disorder were examined in the patient’s serum and in the plasma removed by plasmapheresis. The free concentration was similar in both the plasma removed and in the patient’s serum. Plasmapheresis did not significantly alter the serum concentration of phenytoin and dosage adjustments of phenytoin are therefore unnecessary when single volume exchanges are performed . These authors question the value of plasmapheresis and single volume exchange as an appropriate therapeutic intervention for phenytoin overdose.