`

`D.J. Greenblatt
`
`Drug-Drug Noninteractions
`
`potential DDIs is very large because polypharmacy is so
`common in clinical practice. It is unrealistic to expect
`that a clinical study can be conducted to evaluate each
`of these possibilities. Clinical DDI studies that meet reg(cid:173)
`ulatory standards are expensive, time-consuming, and(cid:173)
`most importantly-involve some risk to volunteer sub(cid:173)
`jects that participate in the studies. Institutional Review
`Boards serve to assure that the risk is low and acceptable,
`but the risk is non -zero nonetheless. Given the time, cost,
`and risk of clinical DDI studies, a "filtering" mechanism is
`needed to identify those drug combinations having high(cid:173)
`est priority for a DDI study to either confirm or exclude a
`clinically important DDI.
`A filtering process that is generally accepted by the sci(cid:173)
`entific community and the FDA is the in vitro metabolic
`model based on human liver microsomes [3,12-1S]. The
`model establishes the transformation of a specific sub(cid:173)
`strate to its principal metabolites by the liver microsomal
`enzyme preparation. The effect of a specific candidate in(cid:173)
`hibitor on that transformation process can be quantita(cid:173)
`tively characterized by either an in vitro inhibition con(cid:173)
`stant (Kd or a 50% inhibitory concentration (IC so ). If
`Cmax represents the maximum plasma concentration of
`the inhibitor attained in vivo with the highest recom(cid:173)
`mended therapeutic doses, a ratio of Cmax divided by Ki
`or IC so can be used to roughly forecast the possibility of a
`clinical DDI. The available predictive models are far from
`perfect [19-26], and regulatory guidelines accordingly are
`very conservative. Current guidelines state that a Cmax/Ki
`or Cmax/ICso ratio less than 0.1 indicates that a DDIis "un(cid:173)
`likely," while greater than 10.0 indicates "probable." For
`the in-between range (0.1 to 10.0), a DDIis deemed "pos(cid:173)
`sible." The boundaries are arbitrary and the range broad,
`but the guidelines do allow targeting of clinical resources
`to the "possible" range. When ratios are less than 0.1,
`studies are generally not needed. Ratios greater than 10.0
`indicate a high enough probability of a DDI that the drug
`combination may actually be prohibited through labeling
`restrictions.
`In this issue, Schwartz and associates [27] report a
`clinical DDI study in which laropiprant-the candidate
`drug under development by the sponsor (Merck)-was
`evaluated as an inhibitory "perpetrator," and rosiglita(cid:173)
`zone served as the substrate "victim." The study serves
`to define the DDI interaction potential between these
`specific drug pairs, but the outcome can be logically
`extended to other substrate victims that, like rosigli(cid:173)
`tazone, are metabolized mainly by the specific Cy(cid:173)
`tochrome P450-2CS (CYP2CS) isoform. As such, rosigilla(cid:173)
`zone is termed an "index" or "probe" substrate for
`CYP2CS.
`The study outcome demonstrated no interaction be(cid:173)
`tween laropiprant and rosiglitazone-a drug-drug nonin-
`
`teraction. This is the hoped-for result, and good news for
`the sponsor. The product label for laropiprant can explic(cid:173)
`itly cite the pharmacokinetic noninteraction between the
`two drugs, and further assure that laropiprant is unlikely
`to inhibit the clearance of other drugs that are substrates
`for CYP2CS.
`Reassuring as a definitive noninteraction study may be,
`there is residual discomfort about whether the study re(cid:173)
`ally needed to be done in the first place. Was the possi(cid:173)
`bility of a clinically important DDI real enough to war(cid:173)
`rant the dollar cost, and the low and acceptable-but still
`non-zero-human subjects risk of a clinical DDI study?
`Invoking the in vitro filter criteria, Schwartz et al. [27]
`state that laropiprant is a moderate in vitro inhibitor of
`CYP2CS, with an IC so in the range of approximately 6.5
`micromolar. No reference is cited to support this critical
`piece of information, and readers have no way to evalu(cid:173)
`ate the validity of the stated IC so value. In any case, if the
`stated IC so is accepted as valid, their reference 15 reports
`a mean laropiprant steady-state Cmax of 2.1 micromolar
`at a dose of 30 mg per day, and 3.9 micromolar at 60 mg
`per day [2S]. The corresponding Cmax/ICso ratios are 0.32
`and 0.60. At 40 mg, the ratio per day is likely to fall be(cid:173)
`tween 0.32 and 0.60, indicating-by FDA criteria-that a
`DDI is "possible." This is reasonable rationale and justifi(cid:173)
`cation for moving forward with a clinical DDI study. It is
`also very likely that the sponsor considered consequences
`of not conducting the clinical study. With a DDI deemed
`"possible" based on in vitro data and not definitively ruled
`out in a clinical trial, the FDA could impose labeling to the
`effect that DDIs are possible, have not been evaluated in
`clinical studies, and that coadministration of laropiprant
`and CYP2CS substrate drugs is either prohibited-or un(cid:173)
`dertaken only with special caution and monitoring-until
`clinical data became available. Such labeling would con(cid:173)
`stitute a restriction of clinical use, and might put the
`sponsor at a competitive disadvantage in the market(cid:173)
`place. All things considered, the sponsor elected to pro(cid:173)
`ceed with the clinical DDI study, as reported by Schwartz
`et al. [27].
`A word about biostatistical analysis: For the Schwartz et al.
`[27] study, statistics are not really needed. No sane per(cid:173)
`son looking at their Figure 2 could possibly argue that
`there is remotest evidence of a DDI. The statisticians can
`sit this one out. But when a study does show a change
`in plasma levels of the substrate victim due to coad(cid:173)
`ministration of the precursor, the scientific and health
`care communities need answers to the following ques(cid:173)
`tions: How big is the DDI? Could it have happened by
`chance? Is the DDI of possible or actual clinical impor(cid:173)
`tance? We turn to biostatisticians to help us come up
`with the answers (though not to answer the questions for
`us).
`
`Cardiovascular Therapeutics 27 (20091226-229 © 2009 Blackwell Publishing Ltd
`
`227
`
`Teva Pharmaceuticals USA, Inc. v. Corcept Therapeutics, Inc.
`PGR2019-00048
`Corcept Ex. 2051, Page 2
`
`

`

`Drug-Drug Noninteractions
`
`D.J. Greenblatt
`
`Unfortunately, FDA guidelines for statistical analysis of
`DDI studies obfuscate and confuse-rather than clarify
`and illuminate-the biomedical phenomenon that we are
`hoping to understand through the study outcome [29].
`The FDA demands that DDI studies be treated as bioe(cid:173)
`quivalence studies-which they most certainly are not.
`Data manipulations, such as logarithmic transformation
`and calculations of geometric or harmonic means, distort
`the real central tendencies expressed as arithmetic means
`of untransformed values. If manuscripts on DDI studies
`submitted to scientific or medical journals simply trans(cid:173)
`plant the FDA-mandated statistical analysis from the reg(cid:173)
`ulatory report, the community of scientists and clinicians
`reading the journal may end up confused rather than en(cid:173)
`lightened about the DDI. As for statistical significance of
`an apparent DDI, the straightforward, transparent, and
`unarguable answer comes from a nonparametric equiv(cid:173)
`alent of Student's t-test, yielding the probability that the
`observed difference could have happened by chance. Fi(cid:173)
`nally we have the clinical importance of a DDI-an issue
`that no statistical method can resolve out of context. Is
`the change in plasma levels of the victim substrate drug
`caused by the DDI sufficiently large to make a clinical dif(cid:173)
`ference, and require some sort of corrective action? Ex(cid:173)
`amples would be: a need for increased monitoring or re(cid:173)
`duced dosage of the substrate, or a drug toxicity hazard
`warranting prohibition of the drug combination. This is
`not a matter of statistics-a small but statistically signifi(cid:173)
`cant DDI may be of no therapeutic importance and pose
`no hazard of drug toxicity. What is needed is an under(cid:173)
`standing of the concentration-response or dose-response
`relationship of the victim drug. With that knowledge, a
`clinical judgment can be made as to whether the effect of
`the DDI is large enough to change the response to the
`victim drug.
`With passing years we have learned that DDIs in
`general have received too much "hype." In an era of
`polypharmacy, the number of concurrently-administered
`drug pairs that might interact is huge, yet the actual
`prevalence of significance DDIs is very small [30-32].
`Nonetheless DDI evaluation is now a permanent piece
`of the drug development process. Drug-drug noninterac(cid:173)
`tion studies provide biomedical and public health infor(cid:173)
`mation as important as the studies with positive results.
`Among the "population" of all DDI studies, too few neg(cid:173)
`ative results means that our filter is too stringent, and we
`probably are failing to conduct some studies that would
`be positive. On the other hand, an excessive number of
`noninteraction results imply that our filter is identifying
`drug combinations that are not realistic candidates for a
`clinical DDI. Filtering mechanisms to predict clinical DDIs
`are imperfect [19-26], and require ongoing refinement to
`improve accuracy.
`
`Conflict of Interest
`
`The authors declare no conflict of interests.
`
`References
`
`1. Farkas D, Shader RI, von Moltke LL, Greenblatt DJ.
`Mechanisms and consequences of drug-drug interactions.
`In: Gad SC, editor. Preclinical Development Handbook:
`AD ME and Biopharmaceutical Properties. Philadelphia:
`Wiley, 2008; 879-917.
`2. Greenblatt DJ, von Moltke LL. Drug-Drug Interactions:
`clinical perspectives. In: Rodrigues AD, editor. Drug-Drug
`Interactions, 2nd ed., New York: Informa Healthcare,
`2008; 643-664.
`3. Lin JH, Lu AY. Inhibition and induction of cytochrome
`P450 and the clinical implications. Clin Pharmacokinet
`1998;35:361-390.
`4. Pelkonen 0, Turpeinen M, Hakkola J, Honkakoski P,
`Hukkanen J, Raunio H. Inhibition and induction of
`human cytochrome P450 enzymes: current status. Arch
`ToxicoI2008;82:667-715.
`5. Honig PK, Wortham DC, Zamani 1<, Conner DP, Mullin
`JC, Cantilena LR. Terfenadine-ketoconazole interaction:
`pharmacokinetic and electrocardiographic consequences.
`Journal of American Medical Association
`1993;269: 1513-1518.
`6. von Moltke LL, Greenblatt DJ, Duan SX, Harmatz JS,
`Shader RI. In vitro prediction of the terfenadine(cid:173)
`ketoconazole pharmacokinetic interaction. Journal of
`Clinical Pharmacology 1994;34: 1222-1227.
`7. Woosley RL, Chen Y, Freiman JP, Gillis RA. Mechanism
`of the cardiotoxic actions of terfenadine. JAMA
`1993;269: 1532-1536.
`8. Crumb WJ, Wible B, Arnold DJ, Payne JP, Brown AM.
`Blockade of multiple human cardiac potassium currents
`by the antihistamine terfenadine: possible mechanism for
`terfenadine-associated cardiotoxicity. Molecular
`Pharmacology 1995;47:181-190.
`9. Rampe D, Wible B, Brown AM, Dage RC. Effects of
`terfenadine and its metabolites on a delayed rectifier K +
`channel cloned from human heart. Molecular Pharmacology
`1993;44: 1240-1245.
`10. Mathews DR, McNutt B, Okerholm R, Flicker M, McBride
`G. Torsades de pointes occurring in association with
`terfenadine use. JAMA 1991;266:2375-2376.
`11. Monahan BP, Ferguson CL, Killeavy ES, Lloyd BK, Troy
`J, Cantilena LR. Torsades de pointes occurring in
`association with terfenadine use. JAMA
`1990;264:2788-2790.
`12. Bjornsson TD, Callaghan JT, Einolf HJ, et al. The conduct
`of in vitro and in vivo drug-drug interaction studies: a
`Pharmaceutical Research and Manufacturers of America
`(PhRMA) perspective. Drug Metab Dispos
`2003;31:815-832.
`
`228
`
`Cardiovascular Therapeutics 27 (2009) 226-229 © 2009 Blackwell Publishing Ltd
`
`Teva Pharmaceuticals USA, Inc. v. Corcept Therapeutics, Inc.
`PGR2019-00048
`Corcept Ex. 2051, Page 3
`
`

`

`D.J. Greenblatt
`
`Drug-Drug Noninteractions
`
`13. Bjomsson TD, Callaghan JT, Einolf HJ, et al. The conduct
`of in vitro and in vivo drug-drug interaction studies: a
`PhRMA perspective. J Clin Pharmacol 2003;43:
`443-469.
`14. Huang SM, Strong JM, Zhang L, et al. New era in drug
`interaction evaluation: US Food and Drug Administration
`update on CYP enzymes, transporters, and the guidance
`process. J Clin Pharmacol 2008;48:662-670.
`15. Huang SM, Temple R, Throckmorton DC, Lesko LJ. Drug
`interaction studies: study design, data analysis, and
`implications for dosing and labeling. Clin Pharmacol Ther
`2007;81:298-304.
`16. Volak LP, Greenblatt DJ, von Moltke LL. In vitro
`approaches to anticipating clinical drug interactions. In:
`Albert P Li, editor. Drug-Drug Interactions in
`Pharmaceutical Development. Hoboken, NJ: Wiley, 2008;
`75-93.
`17. Venkatakrishnan K, von Moltke LL, Obach RS,
`Greenblatt DJ. Drug metabolism and drug interactions:
`application and clinical value of in vitro models. Curr Drug
`Metab 2003;4:423-459.
`18. Venkatakrishnan K, von Moltke LL, Greenblatt DJ.
`Human drug metabolism and the cytochromes P450:
`application and relevance of in vitro models.
`Journal of Clinical Pharmacology 2001;41:
`1149-1179.
`19. von Moltke LL, Greenblatt DJ, Schmider J, Wright CE,
`Harmatz JS, Shader RI. In vitro approaches to predicting
`drug interactions in vivo. Biochemical Pharmacology
`1998;55:113-122.
`20. Greenblatt DJ, He P, von Moltke LL, Court MH. The CYP3
`family. In: Ioannides C, editor. Cytochrome P450: Role in
`the Metabolism and Toxicology of Drugs and Other
`Xenobiotics. Cambridge (UK): Royal Society of
`Chemistry, 2008; 354-383.
`21. Galetin A, Ito K, Hallifax D, Houston JB. CYP3A4
`substrate selection and substitution in the prediction of
`potential drug-drug interactions. J Pharmacol Bxp Ther
`2005;314: 180-190.
`
`22. Ito K, Brown HS, Houston JB. Database analyses for the
`prediction of in vivo drug-drug interactions from in vitro
`data. Br J Clin Pharmacol 2004;57:473-486.
`23. Yao C, Levy RH. Inhibition-based metabolic drug-drug
`interactions: predictions from in vitro data. J Pharm Sci
`2002;91:1923-1935.
`24. Obach RS, Walsky RL, Venkatakrishnan K, Houston JB,
`Tremaine LM. In vitro cytochrome P450 inhibition data
`and the prediction of drug-drug interactions: qualitative
`relationships, quantitative predictions, and the rank-order
`approach. Clin Pharmacol Ther 2005;78:582-592.
`25. Bachmann KA. Inhibition constants, inhibitor
`concentrations and the prediction of inhibitory drug-drug
`interactions: pitfalls, progress and promise. Curr Drug
`Metab 2006;7:1-14.
`26. Lin JH. Sense and nonsense in the prediction of
`drug-drug interactions. Current Drug Metabolism
`2000;1:305-331.
`27. Schwartz JI, Stroh M, Gao B, et al. Effects of laropiprant,
`a selective prostaglandin D2 receptor 1 antagonist, on the
`pharmacokinetics of rosiglitazone. Cardiovascular
`Therapeutics 2009;27:239-245.
`28. Lai E, Wenning LA, Crumley TM, et al. Pharmacokinetics,
`pharmacodynamics, and safety of a prostaglandin D2
`receptor antagonist. Clin Pharmacol Ther 2008;83:840-847.
`29. Greenblatt DJ. Preparation of scientific reports on
`pharmacokinetic drug interaction studies. Journal of
`Clinical Psychopharmacology 2008;28:369-373.
`30. Bergk V, Gasse C, Rothenbacher D, Loew M, Brenner H,
`Haefeli WE. Drug interactions in primary care: impact of a
`new algorithm on risk determination. Clin Pharmacol Ther
`2004;76:85-96.
`31. Glintborg B, Andersen SE, Dalhoff K. Drug-drug
`interactions among recently hospitalised patients(cid:173)
`frequent but mostly clinically insignificant. Bur J Clin
`PharmacoI2005;61:675-681.
`32. DeVane CL. Antidepressant-drug interactions are
`potentially but rarely clinically significant.
`Neuropsychopharmacology 2006;31: 1594-1604.
`
`Cardiovascular Therapeutics 27 (2009) 226-229 © 2009 Blackwell Publishing Ltd
`
`229
`
`Teva Pharmaceuticals USA, Inc. v. Corcept Therapeutics, Inc.
`PGR2019-00048
`Corcept Ex. 2051, Page 4
`
`