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 Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Crean, S. M. Articles by Dematto, C. A. Search for Related Content PubMed PubMed Citation Articles by Crean, S. M. Articles by Dematto, C. A. Related Articles Related Reply Related Article

Reliable Evidence of Safety and Efficacy of Elderly Patients in Randomized Clinical Trials Still Needed

United BioSource Corp, Medford, MA

To the Editor:

In the April 1, 2007, issue of the Journal of Clinical Oncology, Kumar et al¹ present "a systematic review to access the accurate participation of elderly patients" in phase III randomized trials. They correctly assert that while elderly patients shoulder a disproportionate burden of cancer, they remain under-represented in the randomized clinical trials offered in their communities by National Cancer Institute cooperative groups. The research materials consisted of protocols and publications from all completed phase III randomized clinical trials (RCTs; n = 345 studies) conducted by five National Cancer Institute cooperative groups from 1955 to 2000. They found only one study (0.29%) that exclusively enrolled elderly patients, 5% that excluded patients 70 years and older, and the rest a mix of young and older patients in which the elderly were rarely described as a demographic proportion (6.3%) and never reported in terms of primary end points.

Kumar et al's findings underscore how study-level data as a research source provide compelling evidence for a global trend such as selection bias. Although this article falls short of a properly reported systematic review (many parts of the protocol were missing), it seems valid to conclude that the elderly are routinely underserved by publicly funded cancer research programs and under-reported as a group in subsequent publications.

Unfortunately, Kumar et al then inappropriately cobble together patient-level outcomes on the very population that was earlier described as missing or admixed. With no stratified populations to work with, Kumar et al used instead those few trials (n = 15) that report at least 40% enrollees 65 years and older as the elderly cohort and compared survival, event-free survival, and treatment related mortality outcomes with outcomes reported in the remainder of the studies. Astonishingly, Kumar et al conclude that the similarity in these two groups show that "the enrollment of elderly in experimental RCTs is not associated with increased harm to this patient population." This statement is indefensible from the data presented. Kumar et al state that "the exclusion of elderly patients from clinical trials" may be due to "biased decision making by physicians...because of perceived unfavorable benefit/tolerance of experimental treatments." Indeed, the study-level findings support the assumption of age as a risk factor: 17 studies were closed to patients older than 70 years, and the one study reserved for the elderly was for tamoxifen, which the authors describe as "one of the safest drugs in our treatment armamentarium." Therefore, the similarity between groups seems more likely due to patient channeling of older patients to safer regimens and not due to the vigorous contributions of those few elderly patients allowed to participate in mixed-age trials. Since the primary outcomes are not stratified by age group, no conclusions can be made from populations where age is associated with perceived treatment intolerance.

In their effort to go "beyond the work undertaken by earlier investigators" conducted on "the issue of participation," Kumar et al have attempted to describe actual outcomes from an enriched age cohort, their "next best possible scenario." This admixed, confounded population is an insufficient proxy for examining the true risk factor of elderly age. In barreling toward an oversimplified, predefined assertion of "no harm associated in the elderly," the author's conclusions are reckless, and if taken as actual truth, could result in many elderly being seriously injured in trials of toxic agents. A blanket statement about "no associated harm" seems invalid at best and callous at worst, when the analysis is hopelessly confounded and the only measure for treatment intolerance is death.

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The author(s) indicated no potential conflicts of interest.

REFERENCES

1. Kumar A, Soares HP, Balducci L, et al: Treatment tolerance and efficacy in geriatric oncology: A systematic review of phase III randomized trials conducted by five National Cancer Institute–sponsored cooperative groups. J Clin Oncol 25:1272-1276, 2007[Abstract/Free Full Text]

Related Reply

• In Reply
  Ambuj Kumar and Benjamin Djulbegovic
  JCO 2007 25: 4028 [Full Text]

Related Article

• Treatment Tolerance and Efficacy in Geriatric Oncology: A Systematic Review of Phase III Randomized Trials Conducted by Five National Cancer Institute–Sponsored Cooperative Groups
  Ambuj Kumar, Heloisa P. Soares, Lodovico Balducci, and Benjamin Djulbegovic
  JCO 2007 25: 1272-1276 [Abstract] [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 Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Crean, S. M. Articles by Dematto, C. A. Search for Related Content PubMed PubMed Citation Articles by Crean, S. M. Articles by Dematto, C. A. Related Articles Related Reply Related Article