P310 Application of Receiver Operating Characteristics to Assess Digoxin Drug Interaction Potential

Caroline A. Lee , Pharmacokinetics, Dynamics & Metabolism, Pfizer Inc., San Diego, CA
Joseph Bentz , Drexel University, Philadelphia, PA
Michael O'Connor , Drexel University, Philadelphia, PA
Johan Palm , AstraZeneca, Mölndal, Sweden
Harma Ellens , GlaxoSmithKline, King of Prussia, PA
Krisztina Heredi-Szabo , SOLVO Biotechnology, Szeged, Hungary
Dallas Bednarczyk , Novartis, Cambridge, MA
Mitchell Taub , Boehringer-Ingelheim Pharmaceuticals, Inc., Ridgefield, CT
Elke S. Perloff , BD Biosciences, Woburn, MA
Christoph Funk , Dept of Non-Clin Drug Safety, Hoffmann-La Roche Ltd, Basel, Switzerland
Praveen Balimane , Bristol-Myers Squibb
Laurent Salphati , Genentech Inc., South San Francisco, CA
Ailan Guo , Non-Clin Drug Safety, Hoffmann-La Roche Inc., Nutley, NJ
Lalitha Podila , Drug Metabolism & Pharmacokinetics, Boehringer Ingelheim Pharmaceuticals, Inc., Ridgefield, CT
Imad Hanna , Drug Metabolism and Pharmacokinetics, Novartis Institutes for BioMedical Research Inc., East Hanover, NJ
Cindy Xia , Drug Metabolism and Pharmacokinetics, Millennium Pharmaceuticals, Inc., Cambridge, MA
Libin Li , Absorption Systems LP, Exton, PA
Guangqing Xiao , Biogen Inc, Cambridge, MA
Heleen M. Wortelboer , Pharmacokinetics, TNO Quality of Life, Zeist, Netherlands
Dietmar Weitz , Sanofi-Aventis
Youngeen A. Pak , Drug Disposition, Eli Lilly and Company, Indianapolis, IN
Eric Reyner , Pfizer Global Research & Development, San Diego, CA
Jesse Taur , Eisai
Xiaoyan Chu , Drug Metabolism and Pharmacokinetics, Merck & Co Inc-Merck Rsch Lab, Rahway, NJ
Ulrike Gradhand , Merck KGaA
Mark Warren , Optivia
Ganesh Rajaraman , Life Technologies - CellzDirect, Austin, TX
Lei Zhang , Office of Clinical Pharmacology, U.S. Food and Drug Administration, Silver Spring, MD
Clinically significant drug-drug interactions (DDIs) involving P-glycoprotein (P-gp, MDR1) are typically not as concerning as P450-mediated DDIs.  Regulatory agency guidance criteria for in vitro cut-off values establish whether or not a new chemical entity inhibits P-gp in vitro and may subsequently alter the pharmacokinetics of digoxin in vivo.  The current FDA draft publication and guidelines provided by the International Transporter Consortium (ITC) whitepaper recommend that a clinical DDI study be considered for a new chemical entity when the maximum concentration at steady state (I1) divided by its in vitro P-gp inhibitory potency (IC50) is ≥ 0.1 or when the nominal gut concentration (I2) divided by its IC50 is ≥ 10.  Because the in vitro IC50 value is fundamental to the calculation, an appropriate determination of the IC50 is critical to avoid false negative or false positive predictions. 

The objective was to determine in vitro P-gp IC50 values for sixteen compounds and to assess the extent of variability.  Twenty-two pharmaceutical and contract research laboratories collaborated to generate IC50 data using four in vitro cell systems (Caco-2, MDR1-MDCKII, MDR1- LLC-PK1 and MDR1-expressing inside-out membrane vesicles); six equations in the form of efflux ratio, unidirectional flux, or net secretory flux; and multiple commercial nonlinear regression software packages.  P-gp inhibition data for each of the compounds were fitted to several logistic and nonlinear regression analyses.  Inhibition variability was assessed separately for each in vitro system and equations, and together.  In general, MDR1-expressing vesicles exhibited IC50 variability ranging from 2- to 38-fold; variability in the cell-based systems was larger, ranging from 2- to 120-fold.  Much of the variability in the cells was due to intrinsic variability between the cells, even within the same cell line.  Receiver Operating Characteristic (ROC) analysis was applied to determine the appropriate in vitro cutoff values for I1/IC50 and I2/IC50 based on the ROC area under the curve (AUC’) to minimize false negative predictions under two conditions 1) all data collated together to determine “universal” cutoff values and 2) individual ROC analysis for each laboratory to determine in vitro cutoff values to define digoxin DDI potential.  Recommendations will be discussed regarding which cell lines, equations, and calculation methodologies are more reliable for predicting digoxin-related DDIs.