Search

This is a summary report of the conference on “Analytical Methods Validation: Bioavailability, Bioequivalence and Pharmacokinetic Studies.” The conference was held from December 3 to 5,1990, in the Washington, D.C., area and was sponsored by the American Association of Pharmaceutical Scientists, the U.S. Food and Drug Administration, Federation International Pharmaceutique, Health Protection Branch (Canada), and the Association of Official Analytical Chemists. The report presents our assessment of the major agreements and issues discussed at the conference. The report is also intended to provide guiding principles for validation of analytical methods used in bioavailability, bioequivalence, and pharmacokinetics studies in humans and animals. The objectives of the conference were as follows: (1) to reach a consensus on what should be required in analytical methods validation and the procedures to establish validation; (2) to determine processes of application of the validation procedures in bioavailability, bioequivalence, and pharmacokinetics studies; and (3) to develop a report on analytical methods validation that may be referred to in developing future formal guidelines. Acceptable standards for documenting and validating analytical methods with regard to processes, parameters, or data treatments are discussed because of their importance in assessing pharmacokinetic, bioavailability, and bioequivalence studies. Other topics that were considered essential in the conduct of pharmacokinetic studies or in establishing bioequivalency criteria, including measurement of drug metabolites and stereoselective determinations, are also discussed.

The rationale for using the logarithmic transformation on concentration-dependent pharmacokinetic parameters a priori is presented. This rationale is based on theoretical pharmacokinetic and statistical grounds, but is also applicable to the practice of physicians in dealing with variations of drug treatment within and between patients. The implications of the transformation on data analysis, specifically analysis of variance, and estimation and inference from the analysis as it pertains to bioequivalence decisions are explored. Implementation of the transformation is shown, with an example of two perphenazine formulations in a single-dose crossover study. It is concluded that the transformation has to be accepted on theoretical grounds because sample sizes are too small in bioequivalence studies and too susceptible to extreme values to state with any certainty the actual distribution of pharmacokinetic parameters or their differences within a subject.

The bioavailability of three isoniazid formulations was assessed using a procedure specific for the free drug. Nine human volunteers, all slow acetylators, were each given 4 × 100 mg of isoniazid of three different tablet formulations at weekly intervals; the plasma drug levels were measured at different times during the first 24 hr. No significant differences ( p > 0.05) were detected among the three products as to relative bioavailability, peak plasma concentrations, C max , and the time of C max , t max -Analysis of variance of the pharmacokinetic parameters obtained according to a one-compartment open model did not demonstrate any significant formulation or time effect but revealed a significant intersubject variation in all parameters involved.

A sensitive and specific high‐performance liquid chromatographic procedure for the determination of chlorpropamide or tolbutamide in plasma in the presence of their metabolites is described. The ether extract of acidified plasma is redissolved in the mobile phase, 17% acetonitrile in 0.05 M aqueous ammonium formate, and chromatographed on a reverse‐phase column on a high‐performance liquid chromatograph fitted with a UV absorbance detector. Quantitation of plasma samples containing less than 0.5 μ g/ml of chlorpropamide and 5 μ g/ml of tolbutamide is reported, using these drugs as mutual internal standards. The retention times of the metabolites are such that they do not interfere in the procedure. The assay method was tested in a human volunteer with both drugs and found suitable for single‐dose pharmacokinetic studies.

A novel method for the quantitative estimation of warfarin in plasma is described. Plasma containing warfarin to which a known amount of phenylbutazone is added as internal standard is acidified and extracted with ethylene dichloride. The drug and the internal standard are then back-extracted into alkali which, in turn, is acidified and reextracted with ethylene dichloride. The organic extract, after washing with phosphate buffer (pH 7.2), is evaporated and the evaporated extract is reacted with an ethereal solution of diazomethane (100 μl). The reacted mixture is evaporated and then dissolved in 25 μl of carbon disulfide. Ali-quots (2–3 μl are injected into a gas chromatograph equipped with a flame-ionization detector. The methyl derivatives of warfarin and the internal standard give sharp, well-separated, symmetrical peaks. The method is of sufficient sensitivity to determine plasma levels in humans after single doses (20 mg) of warfarin (sensitivity of 0.25 μg/ml).

A GLC method for the quantitative estimation of phenprocoumon from plasma is described. Plasma containing phenprocoumon, to which a known amount of phenytoin is added as the internal standard, is acidified and extracted with ethylene dichloride. The drug and the internal standard are then back-extracted into alkali, which is acidified and reextracted with ethylene dichloride. The organic extract is evaporated, and the evaporated residue is mixed with 50 mul of trimethylanilinium hydroxide in methanol. Aliquots (1-2 mul) are injected into a gas chromatography equipped with a flame-ionization detector in which the injection port is held at 325 degrees. The methyl derivatives of phenprocoumon and the internal standard give sharp, well-separated, symmetrical peaks. The method is of sufficient sensitivity to determine 0.125 mug/ml of the drug in plasma with a coefficient of variation of 7%.

An improved GLC procedure was developed for the simultaneous determination of phenytoin and its metabolite, 5‐( p ‐hydroxyphenyl)‐5‐phenylhydantoin, in plasma and urine following enzyme hydrolysis. After extraction, the drug, the metabolite, and the internal standard, 5‐( p ‐methylphenyl)‐5‐phenylhydantoin, are measured by GLC with flame‐ionization detection as their respective methyl derivatives following flash‐heater methylation with tri‐methylanilinium hydroxide. The drug and metabolite give well‐resolved symmetrical peaks on a phenyl methyl silicone column, and the method has a sensitivity of 150 ng/ml of phenytoin and 125 ng/ml of the metabolite. GLC–mass spectral evidence is presented for the formation and intact determination of methyl derivatives of the drug, its metabolite, and the internal standard.

Dissolution studies of eight different lots of acetaminophen tablets were investigated by means of three simple dissolution procedures. None of the methods provided complete correlation with physiological availability of acetaminophen when estimated from blood or urine profiles. Two lots of tablets, which were several years old, released the drug more slowly in vitro than more recently purchased lots of the same brands. All of the recently manufactured products required less than 15 min. for 50% dissolution; this time appears to be a reasonable limit for tablet formulations of acetaminophen.

A quantitative method for the gamma-oxo metabolite of phenylbutazone from plasma is described. The procedure involved an ethylene dichloride extraction of acidified plasma to which an internal standard, acenocoumarol, had been added. The extracted gamma-oxo metabolite and the internal standard were methylated and analyzed by GLC. Determination of 0.25 microgram of gamma-oxo metabolite/ml with a relative standard deviation of 6.5% was accomplished.

A GLC method for qualitative and quantitative determination of 3,4-methylenedioxyamphetamine in biological fluids is described. After extraction of the drug and the internal standard, 4-methoxyamphetamine, from plasma and urine, the solvent is evaporated and the residue is mixed with 20 mul of freshly distilled ether. Aliquots (1-2 mul) then are injected into the gas chromatograph. Both the drug and the internal standard give well-separated symmetrical peaks. Flame-ionization detection allows concentrations of 0.125 mug of 3,4-methylenedioxyamphetamine from plasma to be determined with a precision of 3.16%. The method is applicable to the estimation of 4-methoxyamphetamine, employing 3,4-methylenedioxyamphetamine as the internal standard. For identification purposes, various derivatives such as Schiff bases, isothiocyanates, trifluoroacetates, and heptafluorobutyrates of both compounds were formed following extraction from urine and their GLC behavior was investigated. Derivative formation was confirmed by GLC-mass spectrometry. Mass spectral data for these derivatives are presented.

For the development of radioimmunoassay procedures for chlorpromazine and its active metabolites, three chlorpromazine haptens, 7-(or 8-)(3-carboxypropionyl)chlorpromazine, N -(3-carboxypropionyl)desmethylchlorpromazine, and N -(2-carboxyethyl)desmethylchlorpromazine, were synthesized and characterized by GLC–mass spectrometry, PMR spectrometry, and IR spectrophotometry. Each hapten was coupled to bovine serum albumin, and the number of hapten residues per mole of bovine serum albumin was calculated by UV spectrophotometric methods. Antibodies to each hapten–protein conjugate were obtained in rabbits, and titers of the antiserums were checked by evaluating their binding characteristics to tritiated chlorpromazine.

A recently described GLC method for the determination of phenylbutazone and its major metabolite, oxyphenbuta- zone [1 -phenyl-2- (p -hydroxyphenyl) -3,5-dioxo-4-n -butylpyrazoli- dine], in plasma was extended to the estimation of the second major metabolite, γ-hydroxyphenylbutazone [l,2-diphenyl-3,5-dioxo-4-(3-hydroxybutyl)pyrazolidine]. The method is sensitive to 1.0 μg/ml.

An electron-capture GLC procedure for determination of plasma ephedrine concentrations is described. The procedure is capable of determining 2 ng/ml of ephedrine and is adequate for following profiles after 25-mg single doses. Pentane extraction of the drug and the internal standard and formation of the N-pentafluorobenzoyl derivatives were followed by GLC. Plasma ephedrine concentrations following a 24-mg dose of ephedrine hydrochloride to a human volunteer are presented. Formation of N-trifluoroacetyl, N-pentafluoropropionyl, N-heptafluorobutryl, and N-pentafluorobenzoyl derivatives and their GLC-mass spectrometric identification are discussed together with comparative electron-capture sensitivities of these derivatives toward a nickel-63 detector. The detection of the N-pentafluorobenzoyl derivative of ephedrine is at least 100-fold greater in sensitivity than detection of the N-trifluoroacetyl derivative.

Sensitive specific methods are described for the determination of phenylbutazone and its metabolite oxyphenbutazone from the same plasma sample. The sample, to which an internal standard 5-(4-hydroxyphenyl)-5-phenylhydantoin (Standard II) is added, is first extracted with ether to remove interfering substances and then with n-heptane under acidic conditions to separate phenylbutazone, which is determined on a gas chromatograph by flash methylation (310°) with trimethylanilinium hydroxide using 1-(o-chlorophenyl)-1-(p-chlorophenyl)-2,2,2-trichloroethane as the external standard (Standard I). The aqueous acidic residue from which the phenylbutazone has been selectively removed is shaken again with ether to extract the oxyphenbutazone, which is analyzed with a different GLC system with flash methylation (310°) against Standard II. The methods are of sufficient sensitivity to determine plasma levels in humans after a 200-mg dose of phenylbutazone (phenylbutazone, 1 μg/ml; oxyphenbutazone, 0.5 μ/gml).

A catechol metabolite, 5-(3,4-dihydroxyphenyl)-5-phenylhydantoin, and a methylated catechol metabolite, 5-(3-methoxy-4-hydroxyphenyl)-5-phenylhydantoin, were identified as urinary metabolites in humans, monkeys, and dogs following the administration of phenytoin. These metabolites were separated from each other and from other known metabolites of phenytoin as n-butyl derivatives by GLC and positively identified by combined GLC-mass spectrometry.

A GLC procedure was developed for the quantitative estimation of intact chlorpropamide and tolbutamide concentrations in plasma; the drugs are used as mutual internal standards. After extraction of plasma containing the drug and internal standard with toluene, the dried residue is treated with ethereal diazomethane to form the methyl derivatives of tolbutamide and chlorpropamide. Aliquots of the ethereal solution are injected into a gas chromatograph equipped with a glass‐lined injection port and glass column packed with a phenyl methyl silicone fluid (OV‐25) on Chromosorb W, which facilitates the intact determination of the methyl derivatives of the drugs. The response to the flame‐ionization detector was linear over a range of 0.20–25 μg/ml, with a 0.05–μg/ml limit of detectability for both drugs. The method compares favorably with a recently developed high‐pressure liquid chromatographic procedure and is adequate for following blood level profiles of single doses of chlorpropamide (125 mg) and tolbutamide (250 mg). Mass spectral evidence showing that intact sulfonylureas are measured is presented.

The period of time after administration over which blood level measurements are required to obtain a reliable bioavailability comparison of two or more formulations of the same drug was considered by the analysis of bioavailability data taken from the literature. The drugs examined, selected to represent a range of absorption and elimination half-lives, were acetaminophen, aminosalicylic acid, chloramphenicol, chlordiazepoxide, digoxin, isoniazid, phenylbutazone, sulfamethizole, tetracycline, and warfarin. For most drugs, ratios of areas under the curve changed little between the end of the absorption period and the time when blood sampling was terminated. Reliable bioavailability comparisons among different brands of the drugs apparently could have been made by blood sampling over 24 hr or less.

Thin-layer chromatographic systems and spray reagents are described for the rapid differentiation of 10 penicillins. Comparative behavior of 18 other antibiotics under these conditions is presented together with other supporting data.

The USP/NF dissolution test (Method I or rotating-basket method) was applied to different commercial tablet formulations of sulfisoxazole, sulfamethoxazole, and tolbutamide. It was found that the hydrochloric acid solution required for testing the sulfa drug tablets corroded the basket so that it could not be used after 40 hr. of exposure to the acid. Because the dissolution time of a sulfamethoxazole formulation in phosphate buffer (pH 7.2) was close to that obtained using compendial conditions, it is suggested that the official dissolution medium could be replaced, with advantage, by the phosphate buffer. The dissolution times of nine different sulfisoxazole formulations tested in a less acidic medium were not comparable in rank order and magnitude to those obtained with the compendial test. Dissolution times of eight different tolbutamide formulations in phosphate buffer (pH 7.6) were similar in rank order and magnitude to those obtained with the compendial test. It is suggested that this buffer replace the tromethamine [tris-(hydroxymethyl)aminomethane] buffer, currently official, which has a large temperature coefficient of pH.

Procedures are described for the qualitative and quantitative determination of vancomycin hydrochloride by means of spectrophotometry and thin-layer chromatography. The Folin-Ciocalteau reaction which forms the basis of both procedures is also amenable to the estimation of vancomycin factor A, a commonly occurring impurity.

A study was designed to test the influence of surface area on the percutaneous absorption of nitroglycerin from a commercial ointment formulation, using a simple crossover design. On separate occasions, three volunteers were given 16 mg of nitroglycerin (2%) over a 25- and 100-cm2 area. Plasma nitroglycerin concentration was measured at 30, 45, 60, and 90 min using a sensitive capillary GLC-electron-capture detection method capable of quantitating to 150 pg/ml. Plasma concentrations at all times increased at least twofold with the increased surface area; highest observed concentrations were 0.17 and 0.41 ng/ml, respectively. A fourth volunteer received 16 and 32 mg of nitroglycerin over 100 cm2. Doubling the dose increased the 0-90-min AUC by only 76% but caused a 3.5-fold increase in the 90-min plasma concentration. These results suggest that the surface area of application significantly influences the pharmacokinetics of nitroglycerin ointment.