|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Originally published as JCO Early Release 10.1200/JCO.2009.23.8204 on June 22 2009 © 2009 American Society of Clinical Oncology.
Stopping Trials Early for Positive Results: The Need to Know How MuchNew York Oncology Hematology, Hudson, NY To the Editor: It is reassuring that Korn et al1 have been able to show that stopping or early reporting of results in randomized clinical trials by the National Cancer Institute Cooperative Groups has generally not led to the kind of biased results noted by Wilcox et al2 that lead to type I errors. However, Korn et al admit, "Concerns about biased treatment effects as a result of the early stopping are statistically valid." They go on to state that "releasing information early about an effective treatment may be more important than knowing the exact magnitude of the benefit." Here I must respectfully disagree. They nicely elucidate in their section "Choice of Primary End Point" the difficulties raised by the use of surrogate end points and the effect of crossovers to the experimental arm after early stopping of a trial. A prime example of the problem is the Tamoxifen Prevention Trial, which was stopped early because the primary end point—breast cancer incidence—crossed a prespecified boundary.3 On follow-up, after treatment crossovers to the experimental arm, no benefit has yet been demonstrated in overall or breast cancer–specific survival.4 Significant toxicities to otherwise healthy women were documented, including rare cases of uterine cancer and somewhat less rare cases of thromboembolism.3 If the estimate of benefit is biased, how important is it to know that treatment of the overall population (with varied risks) was beneficial compared with no treatment? How are prospective patients supposed to determine the value of treatment? Would not more precise information assist me in helping a 50-year-old woman who is worried about blood clots make such a decision? Sometimes we need a trial to tell us that an experimental arm is better than its comparator. More often we need to know how much better.5 AUTHOR'S DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST The author(s) indicated no potential conflicts of interest. REFERENCES
1. Korn EL, Freidlin B, Mooney M: Stopping or reporting early for positive results in randomized clinical trials: The National Cancer Institute Cooperative Group experience from 1990 to 2005. J Clin Oncol 27:1712–1721, 2009. 2. Wilcox RA, Djulbegovic B, Guyatt GH, et al: Randomized trials in oncology stopped early for benefit. J Clin Oncol 26:18–19, 2008. 3. Fisher B, Costantin JP, Wickerham DL, et al: Tamoxifen for prevention of breast cancer: Report of the National Surgical Adjuvant Breast and Bowel Project P-1 study. J Natl Cancer Inst 90:1371–1388, 1998. 4. Chlebowski RT, Col N, Winer EP, et al: American Society of Clinical Oncology technology assessment of pharmacologic interventions for breast cancer risk reduction including tamoxifen, raloxifene, and aromatase inhibition. J Clin Oncol 20:3228–3343, 2002. 5. Rothman KJ: A show of confidence. N Engl J Med 299:1362–1363, 1978.[Medline]
Related Articles
Related Reply
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2009 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
