Unsuspected poor metabolizer phenotypes of fluoxetine in bioavailability/bioequivalence studies from an indian...

Fluoxetine, belonging to the class of selective serotonin uptake inhibitors, has been extensively used for the treatment of depression and other psychiatry related disorders. Fluoxetine (CAS 59333-67-4) is a substrate of polymorphic cytochrome P450 2D6 isozyme (CYP2D6) leading to the formation of norfluoxetine (CAS 83891-03-06), which is not only active but long lived than the parent in the systemic circulation. Since the parent and metabolite levels are important from a therapeutic perspective, knowledge of phenotypic distribution of the population may be an important consideration.

The aim of the work was to retrospectively evaluate the pharmacokinetic data of fluoxetine and norfluoxetine from several bioavailability/bioequivalence (BA/BE) studies to identify the poor metabolizer (PM) phenotypes from the unsuspected healthy subjects across varied protocol designs, dose sizes and differing formulations.

The pharmacokinetic data of fluoxetine and norfluoxetine were gathered from several BA/BE studies conducted at clinical facilities located at Bangalore and Chennai, India. The BA/BE studies involved open label, two-way randomized crossover designs with two periods and two treatments (reference and test). Blood samples were collected for at least 672 h after fluoxetine dosing and the plasma was analyzed using validated tandem liquid chromatography mass spectrometric assay to determine fluoxetine and norfluoxetine levels. Standard pharmacokinetic parameters were computed using noncompartmental methods. For the purpose of this paper, retrospective evaluation of pharmnacokinetic data from only the reference formulation was considered. The AUC ratio of fluoxetine/norfluoxetine was computed. The individual fluoxetine/norfluoxetine AUC(0-infinity), ratios were plotted in increasing rank order and using outlier test ('T' procedure at 5% level of significance) the subjects were categorized as extensive metabolizer (EM) and PM phenotypes. The unequivocal confirmation of PM phenotypes was obtained by performing linear regression analysis of fluoxetine vs norfluoxetine AUC(0-infinity) values.

Each study was evaluated for the distribution of EM and PM phenotypes of fluoxetine. There were 144 subjects evaluable from four studies, 89.6% (129 out of 144) of which could be categorised as EMs and 10.4% (15 out of 144) as PMs of fluoxetine. The pharmacokinetic parameters were quite distinct between the two phenotypes: (1) PM phenotypes showed much higher exposure (approximately 2.3-fold increase in AUC(0-infinity) and much slower elimination (almost 2-fold increase in elimination half-life) for fluoxetine as compared to EM phenotypes; (2) PM phenotypes showed approximately 0.5-fold lower exposure of norfluoxetine as compared to the EM counter parts; (3) There was no change (approximately 1.2-fold) in the elimination half life of norfluoxetine in EM and PM phenotypes.

Retrospective evaluation of fluoxetine and norfluoxetine pharmacokinetic data demonstrated existence of both PM and EM phenotypes in the Indian population. Based on the overall data (n=144 subjects) there appeared to be 10.4% of PM phenotypes for fluoxetine and/or possibly for other polymorphic CYP2D6 substrates commonly used in this region.