1521-009X/41/11/1972$25.00 http://dx.doi.org/10.1124/dmd.113.054361 DRUG METABOLISM AND DISPOSITION Drug Metab Dispos 41:1972, November 2013 Copyright ª 2013 by The American Society for Pharmacology and Experimental Therapeutics Letter to the Editor Endogenous 4b-Hydroxycholesterol-to-Cholesterol Ratio Is Not a Validated Biomarker for the Assessment of CYP3A Activity Received August 20, 2013; accepted August 23, 2013 Björkhem-Bergman et al. (2013) published a study in healthy adults evaluating endogenous 4b-hydroxycholesterol-to-cholesterol ratio to measure CYP3A induction. The authors compare CYP3A induction fold-changes and determined correlations between 4b-hydroxycholesterol- to-cholesterol ratio and midazolam clearance after rifampicin admin- istration. Midazolam clearance is a validated biomarker to evaluate CYP3A activity. We commend the authors for attempting to find alternative, simple, and cost-effective methods to evaluate CYP3A activity. However, we are concerned that the study results may lead to inappropriate use of the 4b-hydroxycholesterol-to-cholesterol ratio for evaluating CYP3A-mediated drug-drug interactions. A statistically significant, but weak, relationship was reported between 4b-hydroxycholesterol-to-cholesterol ratio and midazolam clearance (coefficient of determination [r 2 ] = 0.29, P , 0.01). The authors state that the ratio “…might be used as a marker to evaluate CYP3A activity at baseline and not only during induction” (Björkhem- Bergman et al., 2013). Although correlation coefficients (r) and/or r 2 values are commonly reported in the literature and used to assume suitability of a cytochrome P450 (P450) probe (Fuhr et al., 2007), r and r 2 values provide limited information. Correlation coefficients describe the strength and direction of an association between independent and dependent variables. Coefficients of determination describe the degree of variability between variables. Most importantly, r and r 2 values are not a measure of predictive performance (Sheiner and Beal, 1981; Bland and Altman, 1986). Additional limitations of r 2 values include the inability to determine if the most appropriate set of independent variables was selected, whether independent variables are causes of changes in the dependent variable, and whether omitted- variable bias exists (Nagelkerke, 1991; Draper and Smith, 1998). Additionally, in a previously published study, the authors compared 4b-hydroxycholesterol to another validated CYP3A probe, quinine, and reported weak r values with statistical significance (r = –0.24 to –0.5, P , 0.05) (Diczfalusy et al., 2008). However, statistically significant r and/or r 2 values do not substantiate that a P450 probe is valid for general use. Correlation coefficients and/or coefficients of determination are often overvalued and result in exaggerated con- clusions. A P450 probe should adhere to validation criteria to be con- sidered appropriate for use (Watkins, 1994; Zaigler et al., 2000; Fuhr et al., 2007). Given the limitations of r and r 2 values in general, as well as in these studies (Diczfalusy et al., 2008; Björkhem-Bergman et al., 2013), we do not believe that the 4b-hydroxycholesterol-to-cholesterol ratio has been validated as a biomarker for the measurement of CYP3A activity. Skaggs School of Pharmacy and Pharmaceutical Sciences, University of California, San Diego, La Jolla, California (J.D.M.); and Bertino Consulting, Schenectady, New York (A.N.N., J.S.B.) JOSEPH D. MA ANNE N. NAFZIGER JOSEPH S. BERTINO, Jr. Authorship Contributions Wrote or contributed to the writing of the manuscript: Ma, Nafziger, Bertino. References Björkhem-Bergman L, Bäckström T, Nylén H, Rönquist-Nii Y, Bredberg E, Andersson TB, Bertilsson L, and Diczfalusy U (2013) Comparison of Endogenous 4b-Hydroxycholesterol with Midazolam as Markers for CYP3A4 Induction by Rifampicin. Drug Metab Dispos 41: 1488–1493. Bland JM and Altman DG (1986) Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1:307–310. Diczfalusy U, Miura J, Roh HK, Mirghani RA, Sayi J, Larsson H, Bodin KG, Allqvist A, Jande M, and Kim JW, et al. (2008) 4Beta-hydroxycholesterol is a new endogenous CYP3A marker: relationship to CYP3A5 genotype, quinine 3-hydroxylation and sex in Koreans, Swedes and Tanzanians. Pharmacogenet Genomics 18:201–208. Draper NR and Smith H (1998) Applied Regression Analysis, Wiley-Interscience, New York. Fuhr U, Jetter A, and Kirchheiner J (2007) Appropriate phenotyping procedures for drug me- tabolizing enzymes and transporters in humans and their simultaneous use in the “cocktail” approach. Clin Pharmacol Ther 81:270–283. Nagelkerke NJD (1991) A note on a general definition of the coefficient of determination. Biometrika 78:691–692. Sheiner LB and Beal SL (1981) Some suggestions for measuring predictive performance. J Pharmacokinet Biopharm 9:503–512. Watkins PB (1994) Noninvasive tests of CYP3A enzymes. Pharmacogenetics 4:171–184. Zaigler M, Tantcheva-Poór I, and Fuhr U (2000) Problems and perspectives of phenotyping for drug-metabolizing enzymes in man. Int J Clin Pharmacol Ther 38: 1–9. Address correspondence to: Dr. Joseph D. Ma, University of California, San Diego, Skaggs School of Pharmacy and Pharmaceutical Sciences, 9500 Gilman Dr. #0714, La Jolla, CA 92093-0714. E-mail: joema@ucsd.edu dx.doi.org/10.1124/dmd.113.054361. 1972 at ASPET Journals on February 6, 2016 dmd.aspetjournals.org Downloaded from