Clinical response in epilepsy in relation to total and free serum levels of phenytoin

The relationships between total and free serum concentrations of phenytoin and the clinical control of seizures were investigated retrospectively in 114 patients. Total phenytoin levels were measured by enzyme-modified immunoassay (EMIT), and the free fraction by ultrafiltration at 37 degrees C using 14C-labelled phenytoin as a tracer. The median free fraction in 188 serum samples was 13.7% (range 8.9-27.0%). The free fraction was greater than 18% in 34 (18.1%) of the serum samples. In all but 5 samples, a likely reason for the elevated free fraction could be determined. The identifiable reasons were commonly hypoalbuminaemia and the presence of liver or renal disease. There was a significant negative correlation between serum albumin level and free fraction of phenytoin (n = 90, r = -0.68, p less than 0.001). The free phenytoin concentration was strongly correlated with the total phenytoin concentration in serum (n = 188, r = 0.94, p less than 0.001). The total phenytoin concentration provided as good an indication of clinical response as the free concentration in 91 patients (85.8% of the patients for whom response could be reliably determined). In the other 15 (14.2%) patients, free phenytoin concentrations were better related to clinical effect. These patients generally had significant reductions in the serum protein binding of phenytoin. The relationship between phenytoin toxicity and free serum concentrations was particularly strong--in 14 patients with toxicity, the total serum concentration of phenytoin was greater than 80 mumol/L in only 42.9% of cases, while the free phenytoin concentration was greater than 8 mumol/L in 85.7% of the cases (p less than 0.05 by chi-square test).