Improved determination of phenobarbital, primidone, and phenytoin by use of a preparative instrument for extraction, followed by gas chromatography.

A microprocessor-controlled automatic extractor, (PREMTM I) is described with which lipophilic components may be extracted from physiological fluids by means of a selective, solid-phase extraction cartridge, and the extracts presented in dry form for subsequent analysis. This instrument has been successfully evaluated for the extraction of phenobarbital, primidone, and phenytoin from human serum before chromatographic analysis. We describe a specific method for determination of these anticonvulsants, involving the PREP I and analysis, after flash-heater methylation, in a gas chromatograph equipped with a nitrogen-sensitive detector. Absolute recoveries ranged from 90 to 102%. Within-day and day-to-day precision (CV) ranged from 4 to 7% for all three drugs in therapeutic concentrations. Comparison of this method with results obtained with a liquid--liquid micro-scale extraction and a similar gas-chromatographic method resulted in correlation coefficients of 0.988 (phenobarbital), 0.978 (primidone), and 0.982 (phenytoin). These drugs in therapeutic concentrations can be simultaneously extracted and measured with confidence.

Micro-scale anticonvulsant assay with use of nitrogen/phosphorus detector and on-column methylation compared with a macro-scale procedure involving flame-ionization detection.

We report a modified micro-scale gas-chromatographic procedure for determining phenobarbital, primidone, and diphenylhydantoin in serum with use of a nitrogen/phosphorus detector. Serum, 20 mul, is extracted into chloroform, the solvent is evaporated, and the residue dissolved in dilute (32 mmol/liter) trimethylanillinium hydroxide, which yields the methyl-derivatives of the drugs upon on-column injection. Within-run precision (CV) for the three drugs was 1.0, 2.1, and 1.4+, respectively, at concentrations of 19.2, 13.2, and 18.9 mg/liter; run-to-run precision at these concentrations was 4.4, 3.8, and 2.8%. Comparison of the 20-mul micro-scale procedure (y) with our conventional, flame-ionization macro-scale procedure (x), in which on-column methylation is used and which requires 1 ml of serum, gave the following respective results: y = 1.01 X -- 0.16 mg/liter, r = 0.992, y = 1.02 X + 0.29 mg/liter, r = 0.973; and y = 0.96x + 0.04 mg/liter, r = 0.996.

Quantitative gas-chromatographic flame-ionization method for chloramphenicol in human serum.

We describe a flame-ionization gas-chromatographic procedure for determination of the potentially toxic antibiotic, chloramphenical, in serum. The serum (500 mul) is extracted into ethyl acetate and nonpolar impurities are subsequently partitioned into hexane. The drug is chromatographed as its bis-trimethylsilyl derivative, with the analog thiamphenicol as the internal standard. Within-run precision (CV) is 4.4% at a serum concentration of 41 mg/liter and 9.2% at a concentration of 5 mg/liter. Over a six-month period, the run-to-run variation was 5.1% at 40 mg/liter (n = 24). Results by the gas-chromatographic method compared well with those by an established colorimetric procedure; mean concentrations for the comparison samples in the two procedures were 18.4 mg/liter and 17.6 mg/liter, respectively (n = 27), with a coefficient of correlation of 0.998. The gas-chromatographic method is more precise and specific than classical microbiological procedures and is suitable for routine therapeutic monitoring of serum chloramphenicol concentrations.

Gas-chromatographic micro-scale procedure for theophylline, with use of a nitrogen-sensitive detector.

We report a micro-scale procedure for determination of theophylline by use of gas chromatography with nitrogen-sensitive detection. The procedure requires only 10 mul of serum, plasma, or saliva. This sample size makes the determination especially appropriate for monitoring in a pediatric population. With the sample volume used, background interference is equivalent to about 0.1 mg/liter and 0.5 mg of theophylline per liter can easily be measured. The new method correlates well with our earlier flame ionization gas-chromatographic procedure that required 1 ml of sample [Clin. Chem. 21, 1038 (1975)]. The selectiviey of the nitrogen detector allows a simplified extraction procedure. Between-run precision (CV) is 2.8% at a theophylline concentration of 14.8 mg/liter.

Monitoring drug concentrations in a case of combined overdosage with primidone and methsuximide.

We describe a case of fatal overdosage with primidone and methsuximide. During the early phase of the patient's hospital course we found concentrations of methsuximide, N-desmethylmethsuximide, and primidone in serum that far exceeded the usual therapeutic concentrations, as determined by gas-liquid chromatography. Determination of N-desmethylmethsuximide in peritoneal fluid demonstrated concentrations comparable to those in serum. This led to the therapeutic decision to manage the patient by dialysis. Subsequently, serum samples collected during the course of hospitalization were analyzed quantitatively by gas-liquid chromatography for methsuximide, N-desmethylmethsuximide, primidone, phenobarbital, and diphenylhydantoin. Selected serum specimens were also analyzed by gas chromatography-mass spectrometry, and N-methyl-2-hydroxymethyl-2-phenylsuccinimide, a metabolite of methsuximide not previously described in human serum, was identified by analysis of its mass spectrum.

Therapeutic monitoring of anticonvulsant drugs: gas-chromatographic simultaneous determination of primidone, phenylethylmalonamide, carbamazepine, and diphenylhydantoin.

We describe a sensitive and precise gas-chromatographic method in which benzylmalonate methylester monoamide is used as the internal standard for the simultaneous determination of primidone, phenylethylmalonamide, carbamazepine, and diphenylhydantoin. The trimethylsilyl derivatives of the anticonvulsants are well separated from each other and from normal serum constituents. The lower limit of detection for each drug is 0.5 mg/liter when 1 ml of serum is analyzed. Within-run precision (CV), established by analysis of 10 replicates, was as follows: primidone (5.4 mg/liter), 2.6%; phenylethylmalonamide (5.5 mg/liter), diphenylhydantoin (6.6 mg/liter), 3.8%; and carbamazepine (10.4 mg/liter), 3.2%. Fifty specimens were analyzed for primidone and 35 for diphenylhydantoin by a standard gas-chromatographic method involving on-column methylation and by the procedure we have developed. The mean value observed for primidone with the on-column alkylation procedure was 9.3 mg/liter and with our procedure was 9.6 mg/liter. When values for our assay were regressed against values for the standard method, the slope of the least-squares line was 0.936, the intercept was 1.00 mg/liter, and r was 0.939. The mean values observed for diphenylhydantoin by on-column methylation and with our procedure were both 12.6 mg/liter. When values for our assay were regressed against the standard method, the slope of the least-squares line was 0.944, the intercept was 0.3 mg/liter, and r was 0.988.

A quantitative gas chromatographic determination of ethosuximide based on N-butylation.

A gas chromatographic procedure has been developed for the routine determination of the anticonvulsant, ethosuximide, in serum. Volatile butyl derivatives for gas chromatography were prepared by the method of R.H. Greeley (J. Chromatogr., 88 (1974) 229) using butyl iodide as the alkylating agent. This procedure yielded stable derivatives with improved chromatographic qualities over those obtained by methylation and proved superior to both on-column alkylation with tetrabutylammonium hydroxide and to silylation. The analysis requires 0.5 ml of serum and the between run precision of the assay at 40 mug/ml was 3.3%.