Temperature effects on the estimation of free levels of phenytoin carbamazepine and phenobarbitone.

Serum samples from patients on phenytoin (PHT), carbamazepine (CBZ), or phenobarbitone (PB) monotherapy were filtered at 15, 25, and 37 degrees C and the free concentrations measured by high-performance liquid chromatography. The mean apparent dissociation constants at each temperature were calculated, and were used to predict free drug levels from a further series of patients' samples in which total drug and albumin concentrations only were known. The correlation coefficients (r) between these predicted free levels and experimental results obtained by analysis of the same samples for PHT, CBZ, or PB were 0.977, 0.968, and 0.998, respectively, at 25 degrees C; at 37 degrees C, the corresponding values of r were 0.975, 0.961, and 0.997, respectively. We then determined free fractions (alpha) of PHT, CBZ, and PB at 25 and 37 degrees C and used these values to derive theoretical target ranges for free levels for each of the three drugs. We discuss the implication of these results for patient care, with special reference to the need to specify temperature and quote the appropriate target range when analyses of free levels of AEDs are carried out.

A micromethod for the estimation of free levels of anticonvulsant drugs in serum.

A micromethod for estimating free levels of phenobarbitone, phenytoin and carbamazepine in patients' sera is described. Serum samples are subjected to a process of ultrafiltration, the filtrates treated with acetonitrile and the drug concentration quantified using high performance liquid chromatography. The stability of free levels in specimens before and after storage is investigated. The method is reproducible and mean recovery exceeds 98.5% showing that there is no significant absorption of drug onto the filters used. There is no interference from other substances normally present in patients' sera and there is a good correlation between results obtained by this method and a fluorescence polarisation immunoassay with correlation coefficient between 0.975 and 0.999. Serum samples can be stored for a lengthy period before ultrafiltration without adverse effects. The relevance of the method to patient care is discussed.

Routine monitoring of carbamazepine and carbamazepine-10,11-epoxide in plasma by high-performance liquid chromatography using 10-methoxycarbamazepine as internal standard.

Carbamazepine and carbamazepine-10,11-epoxide were separated by high-performance liquid chromatography (HPLC) with acetonitrile-water as mobile phase, and detection was effected by UV absorption at 215 nm with a total retention time of less than 10 min. Plasma samples were extracted with dichloromethane and 4 M sodium hydroxide, and 10-methoxy-carbamazepine was added as internal standard. Other commonly used anticonvulsant drugs present in plasma showed no significant interference. The within-batch coefficient of variation for carbamazepine was 4.9% and carbamazepine-10,11-epoxide 5.9%. Between-batch coefficients of variation were 3.7% and 5.3%, respectively. Mean recovery for carbamazepine was 100.2% and for carbamazepine-10,11-epoxide 100.6%. This HPLC method was compared with both an enzyme immunoassay procedure (EMIT) and a gas-liquid chromatographic (GLC) method. Correlation coefficient between HPLC/EMIT for carbamazepine was 0.983, HPLC/GLC carbamazepine 0.988 and HPLC/GLC carbamazepine-10,11-epoxide 0.981.

Valproic acid estimation by enzyme immunoassay.

Blood samples from patients on long-term sodium valproate (Epilim) therapy were analysed for valproic acid by two procedures, enzyme immunoassay (EMIT) and gas-liquid chromatography (GLC). A critical evaluation of the effects of anticoagulants added to specimens was performed in addition to studies on reagent specificity, accuracy, and precision. EMIT valproic acid reagent showed high specificity, and there was a good correlation between results obtained by EMIT and GLC over a wide range of concentrations. The accuracy and precision of EMIT assay was good over the therapeutic range of valproic acid. The presence of the anticoagulant EDTA or of fluoride/oxalate produced a bias towards high results for valproic acid as compared with those obtained from the analysis of serum.

A brief trial of phenytoin therapy for thalamic pain.

Phenytoin was given in full dosage to eight patients with thalamic pain and two others with intractable pain resistant to other forms of treatment. Serum levels were monitored and correlated with dosage levels. Three patients improved markedly, two improved only minimally, two were unchanged, and three were worse. Those patients who had improved noted return of original pain on stopping phenytoin. The results indicate the need for a further study of the drug in thalamic and other chronic pain states.