Transcript FDA: Risk Management Mandate

Lead and
Children’s IQ
Clark D. Carrington, Ph.D., D.A.B.T.
Chemical Hazards Assessment Team
Center for Food Safety and Applied Nutrition
US Food and Drug Administration
Lead (Pb) - Hazards







There is little to no margin of exposure/safety - there is some level of
risk, all be it small, associated with low levels of exposure.
Bioavailability of ingested lead are age specific.
In humans the incidence of lead induced effects (adverse below a blood
Pb level of 10 µg/dL ?) is observable and verifiable.
Placenta presents a minimal barrier to movement of lead from maternal
to fetal circulation.
Food consumption per body mass is greater in infants & younger
children (up to approx. 7 years of age and younger) as compared to
adults.
Infants & younger children absorb lead from the GI tract more readily
than adults by about a factor of four.
Immature status of developing organs - brain, kidneys & liver - more
sensitive to toxic insult. The magnitude of response is dose and organ
system dependent.
Key effect of low blood lead levels is on neurocognitive development.
Prospective studies: estimated change in full scale IQ (and 95% confidence interval) for
increase in blood lead from 10 to 20 µg/dl, using three measures of blood lead in each
study. Pocock et al., BMJ 309:1189-1197, 1994.
Carrington & Bolger, Food. Ad. Contamin. 13:61-76, 1996.
Pb and IQ: Model Uncertainty JECFA, 2000
http://www.inchem.org/documents/jecfa/jecmono/v44jec12.htm
Lanphear (2005) Meta-analysis
Hill Model
105
103
101
99
IQ
97
95
93
91
89
87
85
0
5
10
15
20
25
Blood lead level (ug/dL)
30
35
40
Lanphear (2005) Meta-analysis
Bilinear Model
105
103
101
99
IQ
97
95
93
91
89
87
85
0
5
10
15
20
25
Pb lead level (ug/dL)
30
35
40
Lead Issues

Quantitative Dimensions

Individual Response Magnitude




No Threshold (but that doesn’t mean you have to care)
Uncertainty (Sampling, Model)
No Variability
Epidemiological Studies as the Basis for DoseResponse




No Raw Data (usually)
Association means Causation?
Multivariate Analysis (One Wrong, all Wrong)
Systematic Error