14 BIOLOGICAL EFFECTS OF LOW LEVEL EXPOSURES
Q = a “scale” or “affinity” parameter
s = a “shape” parameter
ED 50 = the dose producing one-half of RmAn example of “collapsing” occurs when Ds > > (1/Q); under these condi
tions, R approximates Rm. In toxicity studies, where doses are relatively
large, it is possible to saturate the longevity hormetic response at the lowest
dose; this occurred with 7 -radiation exposure to male and female mice
(lowest dose: 0.11 rad/day), as well as with hexachlorobenzene feeding to
female rats (lowest dose: 0.32 ppm in diet).25 As noted previously,7 longevity
hormesis is frequently a high-affinity, low-capacity phenomenon; that is, it
is manifested and reaches its maximum effect at relatively low doses. Toxic
ity, on the other hand, frequently acts more like a low affinity, high capacity
phenomenon; that is, it is only manifested at “high” doses, and there is
virtually no limit (excepting death) to the damage it can inflict.
THE LONGEVITY HORMESIS DATABASE
Because of (1) the loose use of the term (longevity) hormesis, (2) the ease
with which longevity enhancement from caloric reduction can be confused
with that from longevity hormesis, and (3) the need to analyze age-specific
mortality rate data using appropriate models and weighting factors, only
those data sets analyzed within the framework specified by Neafsey et
al.,24,25 and meeting the appropriate criteria cited therein, will be considered
authenticated. Authentication, however, should only be construed as imply
ing consistency, as opposed to validation (only when the mechanism
becomes known and can be experimentally established will validation be
possible). The criteria for consistency were:
- a good randomness of scatter of data about the fitted curves (judged by
weighted residual plots and visual inspection of the curve-fits24’25,75 - the computed chi-square values were less than the tabulated value (a =
0 .05)14,24’25’41
It should be noted that the data analyzed by Neafsey et al.24,25 focused
exclusively on systems in which mortality from control populations could be
characterized by a linear Gompertz function; because the purpose of these
two papers by Neafsey et al.24,25 was to lay a foundation for the longevity
hormesis concept vis-a-vis Gompertz mortality analysis, it was decided at
that time to simplify the database as much as reasonably possible.
In addition to data sets meeting the specific criteria, there appear to exist
other data strongly suggesting the presence of longevity horme
sis.5,14,24,25,33,47,76"78 These data will also be collated. Table 1.1 summarizes the
database.