The latest data, published in the December 18th issue of the Journal
of the National Cancer Institute, reopen a question that affects millions
of women worldwide: whether to continue tamoxifen beyond 5 years after
surgery for early-stage breast cancer.
Douglass Tormey, MD, PhD, et al, writing for the US Eastern Cooperative
Oncology Group (ECOG) about a study of tamoxifen (Nolvadex) use extended
beyond 5 years, report very early ECOG data that give the first hint in
favor of more than 5 years' duration of tamoxifen therapy. A certain subset
of women in that study--ie, those who test positive for tumor cells with
estrogen receptors--showed a longer time to disease relapse with more than
5 years of tamoxifen, a statistically significant advantage.
The other newly published paper reports early data from a study of 2
vs 5 of years of tamoxifen by the British Cancer Research Campaign Breast
Cancer Trials Group. These data suggest that the longer duration is more
Editorial Questions 1995 NCI Announcement
In an editorial accompanying these first reports from two major, ongoing
tamoxifen clinical trials, Professor Richard Peto, Oxford University, states
that "it may have been unwise" for a 1995 announcement by the
US National Cancer Institute (NCI) "...to have concluded so definitely
that 5 years of treatment is enough...."
The announcement states that "...all available evidence indicates
that 5 years of tamoxifen is a reasonable standard...." Peto says
that much of the evidence used for that determination was available only
because it appeared to be negative--ie, it suggested that taking tamoxifen
for more than 5 years is more dangerous than beneficial--and had been released
earlier than intended on that account. The ECOG evidence, which Peto views
as "apparently favorable," was unavailable until now precisely
because it is positive [and its release could therefore follow a longer,
natural course under the scientific process].
In his editorial, Peto analyzes and compares the newly published results
and the more mature--but still preliminary, in his view--tamoxifen results
published in the November 6th issue of the Journal of the National Cancer
Institute. That issue contains a paper by the Swedish Breast Cancer Cooperative
Group (2 vs 5 years of tamoxifen; the longer duration showing better results
at this point in follow-up) and a paper by Fisher et al for the US National
Surgical Breast and Bowel Project (NSABP)
B-14 study (no tamoxifen vs 5 years and 5 vs 10 years; 5 years' duration
appearing optimal at this point in follow-up). The NSAPB B-14 study, alluded
to above, was stopped and preliminary data were released earlier than intended:
More recurrences of breast cancer were seen among patients assigned to
10 years of tamoxifen than among those receiving 5 years, making it appear
that, although 5 years is clearly beneficial, the risks associated with
longer therapy outweigh the benefits.
Peto, however, stresses that if the suggested hazard were real, the
ECOG results should likewise show an adverse trend, which is not the case.
"...The ECOG data tend, if anything, to favor more than 5 years of
treatment," he says. But he finds that neither set of data is "statistically
convincing" on its own. According to Peto, as-yet insufficient follow-up
time is the chief reason that it is inappropriate to try to synthesize
the data from the four trials through a formal meta-analysis, "and
the early findings in such trials may be therapeutically misleading."
In their paper on the ECOG study, Tormey et al point out differences in
patient populations among the studies that could explain the apparent discrepancies
No Definitive Conclusion About Long-Term Survival Yet Available
Peto says direct and indirect comparisons suggest an advantage for 5
years of tamoxifen over 2 years, although "a definitive conclusion
about long-term survival may not be possible until at least the year 2000.
But neither direct nor indirect comparisons can yet address the question
of whether substantially more than 5 years of adjuvant tamoxifen treatment
will yield better long-term survival."
Longer follow-up of NSABP B-14, the ECOG study, and a Scottish trial
of 5 years vs longer will help, he says, but trials that are still recruiting
patients need to enroll much larger numbers. If "really large-scale
recruitment" can be achieved before the year 2000, "then [these
studies] will yield preliminary findings in 2005 and reliable findings
in 2010," he claims.
In the meantime, Peto believes that the current results from the four
trials reported in the Journal will likely foster agreement "....that
5 years of tamoxifen is a reasonable standard." He maintains, however,
that "...they should also foster continuing disagreement as to whether
longer treatment is promising...."
He concludes that the scientific process "is frustratingly slow,
but eventually it is reliable, and it needs to be."