Does early-life lead exposure predict adult arrests?
Children with higher lead levels in their blood before birth and in early childhood were arrested more often as young adults, including for violent crimes.
Source
Association of prenatal and childhood blood lead concentrations with criminal arrests in early adulthood
Study at a glance
- Design
- Cohort — Prospective birth cohort (Cincinnati Lead Study) with repeated blood lead tests from pregnancy to age 6.5, linked to county criminal-arrest records in early adulthood; negative binomial regression.
- N
- N=250 · 250 cohort members aged 19 to 24 followed through at least the first six years of life; prenatal blood lead was available for 217 of them.
- Population
- Children born in impoverished, older-housing neighbourhoods of Cincinnati between 1979 and 1984, about 90% African-American, followed into young adulthood.
- Outcome
- Number of criminal arrests after age 18 (total and for violent offences), from Hamilton County records.
Structured fields used in claim comparison tables when every cited study has a complete layer.
What they did
The researchers followed 250 people from the Cincinnati Lead Study, whose mothers were recruited during pregnancy in poor inner-city neighbourhoods with lead-contaminated housing. Blood lead was measured in the mother during pregnancy and in the child about every three months to age 5 and then twice a year to age 6.5. When participants were 19 to 24, the team counted their adult arrests from county records, with coders blind to lead levels, and related arrest rates to lead after adjusting for maternal IQ, sex, socioeconomic status and maternal education.
What they found
About 55% of participants had at least one adult arrest, and there were 800 arrests in total. Each 5 μg/dl increase in prenatal blood lead was linked to a 40% higher rate of total arrests (rate ratio 1.40), and each 5 μg/dl increase in blood lead at age 6 to a 27% higher rate (1.27); average childhood lead was not significantly linked to total arrests. For violent arrests, average childhood lead (1.30) and 6-year lead (1.48) were significant, while prenatal lead was not. Males were arrested far more often, but the lead effect did not statistically differ by sex.
The limits
What it doesn't show
Arrests undercount real criminal behaviour and only one county's records were searched, so some arrests were missed; follow-up after age 18 averaged under five years. As an observational study, it cannot rule out confounding by poverty and other disadvantages that cluster with lead exposure, even after adjustment. Blood lead is only an indirect measure of the dose reaching the brain. The cohort was almost entirely African-American and low-income, so racial or broader population differences could not be tested.
Key terms
- Prospective birth cohort
- A study that enrols people at or before birth and measures exposures before the outcomes happen, which helps establish the time order of cause and effect.
- Rate ratio (RR)
- How many times higher the rate of an event (here, arrests per year) is for one exposure level compared with another.
- Attributable risk
- The absolute difference in event rates between higher- and lower-exposed people; here, extra arrests per year at the 95th versus 5th percentile of blood lead.
- Negative binomial regression
- A regression model for count outcomes, like number of arrests, that allows the counts to be more spread out than a simple Poisson model assumes.
- Confounder
- A factor linked to both the exposure and the outcome, such as family poverty, that can create a misleading association if not controlled.
Flashcards
0 of 11 answers reviewed
Research intelligence for this paper
See its role on concept claims, tensions it is part of, placement history, and related discoveries.
Quiz yourself
What is the key design strength of this study for linking lead to arrests?
Common questions
Why count arrests rather than convictions?
The authors argue arrests are closer in time to the actual behaviour, are recorded more completely in the county database, and are less distorted by plea bargaining and diversion programs than convictions.
Why didn't the authors adjust for the child's IQ?
Lower IQ may be one of the ways lead leads to criminal behaviour, so adjusting for it would remove part of the real effect; a variable on the causal pathway is not a true confounder.
Does this prove lead causes crime?
Not on its own. It is observational, so residual confounding remains possible, but the prospective design, repeated exposure measurements and consistency with animal and neurobehavioural evidence make a causal role plausible.
More on Paediatrics