on bisphenol A
1998 industry scientists began publicizing a
study funded by the Society of the Plastics Industy and the
European Chemical Industry Council purporting to show that vom Saal's
work on low-level
bisphenol A impacts could not be replicated.
study was first released in an industry press conference, in which
the conclusions were made available to the press but data sufficient
to allow independent confirmation of the results were not.
also promoted the study aggressively at the first international
meeting of the Japanese Society for Endocrine Disruption Research
(December 1998; see below).
later it was published in the journal Toxicological
made a great deal of this non-replication, and continues to do so
as of August 2000. Unfortunately, not only was their effort flawed
(see below), but independent scientists have
now replicated vom Saal's work. Thus this episode teaches
industry tactics than science.
initial release touted several parts of industry's work:
followed vom Saal's procedure as closely as possible.
sample size was significantly larger.
addition to a control group and an experimental group, they also
included a positive
control group in which they exposed fetal mice to a single
low level of diethylstilbestrol.
found no impact of bisphenol A treatment and no impact of exposure
concluded that their failure to replicate vom Saal exonerated bisphenol
A, and prominently promoted this conclusion to the press. Industry
lawyer Jim Lamb was even quoted in the British science journal,
Nature, to this effect.
the press release unwittingly damned their effort, at least for
the informed reader, revealing that their failure to replicate was
more likely due to an incompetent laboratory. This is because of
the failure of their positive
control with DES. By acknowledging that their positive control
had failed, they acknowledged that their experiments were invalid.
use positive controls to demonstrate they can perform the experimental
procedure. Positive controls should yield positive results. The
fact that they found no positive impact of DES casts doubt on their
the Toxicological Sciences publication of the result no longer
refers to the DES work as a positive control. And industry scientist
John Ashby (Zeneca) has since published on the industry's bisphenol
A web page arguing
(a PDF file) that DES was an inappropriate positive control. These
appear more like damage control than substance.
debate over vom Saal's work became especially heated at the Japanese
meeting in Kyoto in 1998. Industry held daily press conferences
to influence press coverage. They excluded from these conferences
knowledgeable, independent scientists. In fact, when a contingent
of independent scientists attempted to attend one of the industry
press conferences, the industry rep, Jim Lamb, quickly adjourned
the meeting only a few minutes after it had started.
within the industry delegation also is reported to have made a highly
questionable telephone call from Kyoto to a senior scientist/administrator
at the National Institute of Environmental Health Sciences (USA),
in an effort to suppress data from NIEHS that would have shown that
using DES as a positive control was proper. Industry knew these
data existed and it would have completely undermined their case.
The call had its impact; the data were not released.
2002, data began to emerge indicating that the industry's failure
to replicate arose because of inadvertent estrogenization of the
study's control group of mice. The prostates of the control group
were already enlarged above normal, making further enlargement by
additional estrogen signaling (i.e., by the addition of bisphenol
A) impossible. A 2003
publication by Welshons et al. explores in depth the
potential for false
negatives when background estrogen contamination causes positive
controls to fail.