Down's and Frankowski's Comments on Papers by Andersen and Conolly and Bogen

Paper by Anderson and Conolly

Andersen and Conolly discuss several aspects of liver tumor promotion by TCDD. They posit a model for enzyme induction that has five sets of parameter values one set for each of five regions of the liver.Their model for enzyme induction, equation (1), is as follows:

rate= (basal rate)+[(maximal rate)(Ah-TCDD)n]/[(Ah-TCDD)n + (Kdi)n].

This can be written in the form of a linear ratio, viz:

= (x + b)/(cx + d)

where, in their terminology,

= rate, the rate of transcription of CYP1A m-RNA related to the occupancy of Dioxin Response Elements on DNA by the Ah-TCDD complex;

x is the nth power of the Ah-TCDD complex: x = (Ah-TCDD)n,

a is the sum of the basal and maximal rates: a = basal rate + maximal rate,

b is the basal rate times the nth power of Kdi: b = (basal rate) (Kdi)n ,

c is unity, and

d is the nth power of Kdi: d = (Kdi)n.

This is similar to our model, but represents the effective dose in our model. The similarity would be more complete if , x, c and d remained as above, but the a coefficient was simply the maximal rate, and not the sum of the maximal and basal rates. This would entail rewriting equation (1) as:

rate= (basal rate)+[(maximal rate)- basal rate)(Ah-TCDD)n]/[(Ah-TCDD)n + (Kdi)n].

All the enzymatic equations above represent Michaelis-Menten-type enzymatic processes. Such equations play a central role in their model as well as in ours.


This article incorporates a vast array of data into an ecological analysis of the relationship between environmental exposure to radon and lung cancer mortality. The objective of this extensive analysis was to control for potential confounding in areas such as occupational exposures, smoking , and age at death.

Inevitably, the author could only sketch a brief outline of the extensive statistical and demographic methodologies employed, and provide only a summary of key results and implications. These circumstances may result in an understandable reluctance on the part of some to accept his findings.

Some people, for instance, would want to know the overlap between the '40 and over' group and the '60 and over' group, since a large amount of overlap would vitiate whatever conclusions may be drawn from the similarity of results in the two groups. These concerns have been diminished, however, by the obvious care and concern the author has shown in the compiling, checking, and analyses of his data.

There are always nagging details, and the fundamental problems with ecological analyses remain. Nevertheless, the author's methodology and results appear to confirm and strengthen his earlier finding in 1997 of a hormetic relationship between lung cancer mortality and low level radon exposures.