Environ Corporation Ruston, LA 71270
To the Editor:
Waddell (2003) claimed evidence of "a clear and consistent threshold" for bladder and liver cancer in the large ED01 carcinogenicity study. Waddells method involved graphing the excess percentage of tumors versus the dose expressed as the number of molecules per kg body weight per day, and noting that these graphs showed no increase until the dose increased from 1 molecule up to something on the order of 1020 molecules, and then rose almost vertically from that point.
This method is highly flawed. Clearly, there could be no increased responses in the 19-log dose range below 1019 molecules because there were no doses in that range. The "very steep slopes" indicated by Waddell are a visual artifact resulting from his scale in which only about a 1/4 inch of abscissa represents a tenfold increase in dose. Even a linear dose response will produce graphs like those of Waddell when plotted in the same manner. To illustrate, Figure A shows a straight line that passes through the origin plotted between doses of 1018 and 1019 molecules/kg/day. Figure B
shows the same line plotted using Waddells method. Although Waddells method would conclude that a threshold exists at about 1018 molecules/kg/day, this conclusion is clearly not justified. Waddells method will produce similar graphs whether or not there is a threshold.
|
REFERENCE
Waddell, W. J. (2003). Thresholds of carcinogenicity in the ED01 study. Toxicol. Sci. 72, 158163.
|