This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Desta, Z. Articles by Flockhart, D. A. Search for Related Content PubMed Articles by Desta, Z. Articles by Flockhart, D. A. Related Articles Related Correspondence Social Bookmarking                            What's this?

In Reply

Division of Clinical Pharmacology, Department of Medicine, Indiana University School of Medicine, Indianapolis, IN

We read with interest the comments of Lash et al,¹ on our recent editorial² describing the current status of germline pharmacogenetics of tamoxifen. Our piece was prompted by the article by Schroth et al³ suggesting that patients who inherit nonfunctional alleles of cytochrome P450 2D6 respond less well to tamoxifen treatment. We would like to thank these authors for giving us the opportunity to expand on the data underlying our opinion on the subject.

Our editorial attempted to address a series of outstanding questions that remain in this area, and to highlight the fact that much work remains. Lash et al¹ believe that we should have addressed more directly the key question posed by the data of Schroth et al³: whether tamoxifen response is influenced by CYP2D6 variants at all. To support their argument, they present an analysis of data from five diverse studies, using the departure of the distribution of the relative risks from the relative risk of 1 (null). Each study's plot of relative risk (and 95% CI) was plotted against the inverse normal of its rank percentile. If all the studies were considered to be of equal size and quality, and no weighting was applied, this analysis showed that the relative risks fall in the regression line which intersects the x-axis at zero. The authors concluded that the accumulated evidence is a random sample from a log-normal distribution of relative risks centered on the null. On the basis of this line of evidence they contend that there is no evidence that tamoxifen is inferior in patients who carry CYP2D6 variants associated with reduced or absent function.

We believe that the approach employed by these authors is based on the flawed assumption that it is appropriate to treat data obtained in carefully controlled, randomized, prospective trials (ie, level one evidence) with data obtained by analyzing samples in databanks where there was no prospective plan to collect any of the data analyzed, and where critical confounders such as drug dose, compliance, co-medication data, and stage of disease could not be adjusted for. Simply stated, the possibility of such potentially dangerous confounding is why we regard prospective and randomized trials as the gold standard of evidence. Only one of the studies^4,5 used by Lash et al in their analysis is such a prospective trial. The others are all vulnerable to various degrees of confounding.

In the analysis by Lash et al,¹ the relative risks from multiple reported results are ranked from low to high and they are close to a normal distribution. This is equivalent to averaging all five relative risks. Although such an approach might be useful in studies that are equivalent to each other, this simple approach is not appropriate to summarize data from multiple studies, especially retrospective nonrandomized studies in which patient and breast cancer heterogeneity can be great.

We would contend that the strength of evidence from retrospective studies cannot be equated with that obtained from prospective randomized trials. For example, Goetz et al,^4,5 reported that patients homozygous for the *4 genotype (*4/*4) or taking CYP2D6 inhibitors had a worse recurrence-free time and disease-free survival than those with *1/4 and *1/*1 genotypes or those taking no inhibitors, showing the largest effects among those analyzed by Lash et al.¹ These data are derived from a prospective clinical trial, with a well-defined population of postmenopausal women and substantial (more than 12 years) follow-up, and, therefore, the strength of association and the size of effect seems more compelling in this study than most of the other studies that are retrospective in nature and involve less-characterized patients with breast cancer.

In addition, we believe that these authors omitted a number of recent studies showing associations between CYP2D6 variation and reduced tamoxifen response in a range of settings including metastatic breast cancer,⁶ breast cancer prevention,^7,8 and adjuvant therapy.⁹ Of note, data from a second prospective randomized trial have been presented from the Italian tamoxifen prevention trial.^7,8 Data from these trials are consistent with those from the first prospective, randomized trial presented by Goetz et al,^4,5 and indicate an increased risk of disease recurrence in poor metabolizers of CYP2D6. Consistent with an important relationship between disease-free survival and CYP2D6 activity, this study also showed that patients who carried multiple copies of the CYP2D6 gene, associated with increased CYP2D6 activity, experienced improved disease-free survival (P = .0008).⁸ The analysis performed by Lash et al¹ includes data presented by Wegman et al,^10,11 who reported the opposite effect to those trials mentioned above,^3-9 and Nowell et al,¹² who reported no effect of CYP2D6 variant on tamoxifen response. The reason for this discrepancy is not precisely known, but a higher drop-out rate from tamoxifen in extensive metabolizers who experience more severe hot flashes, as demonstrated recently by Rae et al,¹³ is one potential explanation. In this study no poor metabolizers dropped out of a 297-patient trial with an overall 30% drop-out rate, suggesting that patients who stand to benefit the least stay on tamoxifen the longest, whereas those destined to derive the most benefit also experience the most adverse effects, and drop out most frequently.

Although we agree with Lash et al¹ that much more data are required and that many important questions remain, the key principles of the primacy of randomized, prospectively controlled trial data, and the importance of biologic and mechanistic plausibility should be taken into account in making key therapeutic decisions such as this.

We believe that they arrived at their conclusions through an incomplete and imperfect analysis in which they treated nonrandomized retrospective trials and prospective randomized trials as equivalent. Although we recognize that important mechanistic data may still be forthcoming, the available mechanistic data linking the active metabolites of tamoxifen (endoxifen and probably 4-hydroxytamoxifen) to tamoxifen activity and to CYP2D6 metabolic status, and the clinical data showing that CYP2D6 diminished function is associated with worse outcome would seem to carry substantial biologic plausibility at present. The importance of a plausible mechanism is that it suggests rational hypotheses that can be further tested, in this case by examining trials in other clinical settings such as the metastatic, neoadjuvant, or prevention settings, or by examining the effects of inhibitors of tamoxifen metabolism. Drug interaction studies are likely less robust indicators of the effect of CYP2D6 activity than constitutive genotype because of the effects of incomplete enzymatic inhibition, variable compliance, and incomplete period of co-medication. Overall, we believe that the balance of reliable evidence available at present suggests an important relationship between slow CYP2D6 activity and poor response to tamoxifen.

AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a "U" are those for which no compensation was received; those relationships marked with a "C" were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.

Employment or Leadership Position: None Consultant or Advisory Role: David A. Flockhart, Labcorp, Roche (C) Stock Ownership: None Honoraria: David A. Flockhart, Roche Research Funding: None Expert Testimony: None Other Remuneration: None

ACKNOWLEDGMENTS

Supported in part by National Institute of General Medical Sciences Grants No. 1R01GM078501-01A and U01 GM061373-05. We thank Lang Li, PhD, Biostatistician, Indiana University, and James Rae, MD, University of Michigan, for valuable discussions during the preparation of this reply.

REFERENCES

1. Lash TL, Ahern TP, Cronin-Fenton D, et al: Modification of tamoxifen response: What have we learned? J Clin Oncol 26:1764-1765, 2008[Free Full Text]

2. Desta Z, Flockhart DA: Germline pharmacogenetics of tamoxifen response: Have we learned enough? J Clin Oncol 25:5147-5149, 2007[Free Full Text]

3. Schroth W, Antoniadou L, Fritz P, et al: Breast cancer treatment outcome with adjuvant tamoxifen in relation to patient CYP2D6 and CYP2C19 genotypes. J Clin Oncol 25:5187-5193, 2007[Abstract/Free Full Text]

4. Goetz MP, Rae JM, Suman VJ, et al: Pharmacogenetics of tamoxifen biotransformation is associated with clinical outcomes of efficacy and hot flashes. J Clin Oncol 23:9312-9318, 2005[Abstract/Free Full Text]

5. Goetz MP, Knox SK, Suman VJ, et al: The impact of cytochrome P450 2D6 metabolism in women receiving adjuvant tamoxifen. Breast Cancer Res Treat 101:113-121, 2007[CrossRef][Medline]

6. Lim H-S, Lee HJ, Lee ES, et al: Clinical implications of CYP2D6 genotypes predictive of tamoxifen pharmacokinetics in metastatic breast cancer. J Clin Oncol 25:3837-3845, 2007[Abstract/Free Full Text]

7. Bonanni B, Macis D, Maisonneuve P, et al: Polymorphism in the CYP2D6 tamoxifen-metabolizing gene influences clinical effect but not hot flashes: Data from the Italian tamoxen trial. J Clin Oncol 24:3708-3709, 2006[Free Full Text]

8. Bonanni B, Maisonneuve P, Johansson H, et al: Italy risk stratification based on the CYP2D6 tamoxifen metabolizing gene within the Italian tamoxifen prevention trial. Poster presented at the 30th Annual San Antonio Breast Cancer Symposium, San Antonio, TX, December 13-26, 2007

9. Gonzalez-Santiago S, Zárate R, Haba-Rodríguez J A, et al: CYP2D6*4 polymorphism as blood predictive biomarker of breast cancer relapse in patients receiving adjuvant tamoxifen. J Clin Oncol 25:25s, 2007 (suppl, abstr 590)[CrossRef]

10. Wegman P, Vainikka L, Stal O, et al: Genotype of metabolic enzymes and the benefit of tamoxifen in postmenopausal breast cancer patients. Breast Cancer Res 7:R282-R290, 2005

11. Wegman P, Elingarami S, Carstensen J, et al: Genetic variants of CYP2A5, CYP2D6, SULT1A1, UGT2B15 and tamoxifen response in postmenopausal patients with breast cancer. Breast Cancer Res 9:R7, 2007[CrossRef][Medline]

12. Nowell SA, Ahn J, Rae JM, et al: Association of genetic variation in tamoxifen-metabolizing enzymes with overall survival and recurrence of disease in breast cancer patients. Breast Cancer Res Treat 91:249-258, 2005[CrossRef][Medline]

13. Rae JM, Sikora MJ, Henry NL, et al: Cytrochrome P450 2D6 activity predicts adherence to tamoxifen therapy. Oral presentation at the 30th Annual San Antonio Breast Cancer Symposium, San Antonio, TX, December 13-26, 2007

CiteULike    Complore    Connotea    Del.icio.us    Digg    Facebook    Reddit    Technorati    Twitter    What's this?

Related Correspondence

• Modification of Tamoxifen Response: What Have We Learned?
  Timothy L. Lash, Thomas P. Ahern, Deirdre Cronin-Fenton, Jens Peter Garne, Stephen Hamilton-Dutoit, Marianne Ewertz Kvistgaard, Carol L. Rosenberg, Rebecca A. Silliman, and Henrik Toft Sørensen
  JCO 2008 26: 1764-1765 [Full Text]

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Desta, Z. Articles by Flockhart, D. A. Search for Related Content PubMed Articles by Desta, Z. Articles by Flockhart, D. A. Related Articles Related Correspondence Social Bookmarking                            What's this?