Ej. Calabrese et La. Baldwin, A QUANTITATIVELY-BASED METHODOLOGY FOR THE EVALUATION OF CHEMICAL HORMESIS, Human and ecological risk assessment, 3(4), 1997, pp. 545-554
The recent creation of a chemical hormesis database revealed approxima
tely 350 studies qualitatively judged to show high, moderate, or low e
vidence of hormesis. A quantitative methodology was subsequently devel
oped to more rigorously evaluate an experiment's capacity to demonstra
te hormesis. Criteria were established and assigned point values based
on: 1) the number of experimental doses below the high NOEL (i.e., wi
thin the hermetic zone); 2) experimental determination or estimation o
f the high NOEL; 3) the statistical significance of the stimulatory re
sponse; 4) the magnitude (percentage of control value) of the stimulat
ory response; and 5) the reproducibility of data by other studies. Evi
dence of hormesis was assessed by comparing the summation of point val
ues to point ranges established for six evidence categories: high, mod
erate-high, moderate, low-moderate, low, and no-low. Due to the emphas
is on rigorous study design and statistical analysis criteria, 53% of
the experiments received total scores within the ''low'' evidence cate
gory, while only 17% were ranked in the ''high'' and ''moderate-high''
categories. These results reflect the scheme's strengths in that it r
ewards studies with 1) statistically significant data; 2) multiple dos
es within the hermetic zone (i.e., below the high NOEL); and 3) a high
magnitude of stimulatory response. Future refinements should address
limitations of the methodology (e.g., studies showing low dose stimula
tion with low magnitude of response are penalized). Overall, the prese
nt findings are consistent with the more qualitative judgment that che
mical hormesis appears to be a highly generalizable phenomenon with re
spect to biological model, endpoint, and chemical class.