When crime rates began to drop across the U.S. during the 1990s, city officials and criminologists were thrilled—but baffled. Violent acts, most often committed by young adults, had reached an all-time high at the start of the decade, and there was no sign of a turnaround.
By the close of the ’90s, though, the homicide rate had declined more than 40% throughout the country. Economists and criminologists have since proposed reasons for the unexpected plummet. Some have pointed to an increase in police officers. Others have suggested a rise in the number of offenders put behind bars. Economist and “Freakonomics” coauthor Steven D. Levitt famously hypothesized that the legalization of abortion in 1973 even played a role. Once the Supreme Court decided Roe v. Wade, he argued, fewer unwanted babies grew into disturbed, crime-prone adults two decades later.
But recently, experts have been kicking around another possible player in the crime drop of the ’90s: lead. Cars burning leaded gasoline spewed the heavy metal into the air until 1973, when the Environmental Protection Agency mandated the fuel’s gradual phaseout. Lead-based paint was banned from newly built homes in 1978. Because of these actions, children born in the mid- to late-1970s grew up with less lead in their bodies than children born earlier. As a result, economists argue, kids born in the ’70s reached adulthood in the ’90s with healthier brains and less of a penchant for violence.
Today, the Centers for Disease Control & Prevention considers 5 micrograms per deciliter of lead in a child’s blood to be abnormal. Studies have shown that people who grew up with blood-lead levels at or above this threshold are more likely to have impaired cognition than those who grew up with less lead in their blood. In 1976, the average U.S. resident had a blood-lead level of 16 µg/dL, according to the National Health & Nutrition Examination Survey. By 1991, when there was less lead in the air and in housing, the average had dropped to 3 µg/dL.
As the lead-crime hypothesis gains traction in economics circles, critics are invoking the “correlation does not equal causation” mantra. But scientists argue that there is evidence that lead exposure increases aggression in lab animals. And even though lead, one of the oldest known poisons, affects the brain in a dizzying number of ways, researchers are beginning to tease out some of the mechanisms by which it might trigger violence in humans.
During the 1960s, doctors couldn’t label a child as lead poisoned unless he or she had a blood-lead level of at least 60 µg/dL—CDC’s defined limit at the time. But researchers like University of Pittsburgh psychiatrist Herbert L. Needleman questioned the cut-off value. Surely if 60 µg/dL was toxic, 50 µg/dL couldn’t be completely harmless.
Needleman and others began observing “silent lead poisoning” in children with blood-lead levels below the established limit. Rather than overt physical symptoms like hallucinations and kidney damage, these kids had low IQ scores, attention problems, and antisocial tendencies. As more and more reports of these deficits filtered in, CDC lowered the blood-lead level it deemed acceptable for kids further and further: In 1970, the amount was 40 µg/dL, and by 1991, it was 10 µg/dL.
Some physicians noticed that children exposed to blood-lead levels below 50 µg/dL could also be aggressive or violent. In 1996, Needleman and his group followed up on these anecdotal observations by examining a few hundred 12-year-old boys in the Pittsburgh area. The researchers measured the amount of lead in the boys’ bones with X-ray fluorescence to get an idea of how much of the heavy metal their participants were exposed to during childhood. The boys rated worst by their parents and teachers in terms of aggressive and antisocial behaviors had been exposed to the highest levels of lead (J. Am. Med. Assoc. 1996, DOI: 10.1001/jama.1996.03530290033034).
In 2002, Needleman’s team delved deeper by studying 15-year-old boys who had been arrested and sentenced by the Allegheny County Juvenile Court in Pennsylvania. The kids who had been in trouble with the law had an average bone-lead level of 11 ppm—6% higher than a control group of boys without a history of arrest (Neurotoxicol. Teratol. 2002, DOI: 10.1016/s0892-0362(02)00269-6).
Looking for explanations of the ’90s crime drop in the U.S., economists and crime experts latched onto these and other epidemiology studies. “We saw these correlations for individuals and thought, ‘If that’s true, we should see it at an aggregate level, for the whole population,’ ” says Paul B. Stretesky, a criminologist at the University of Colorado, Denver. In 2001, while at Colorado State University, Stretesky looked at data for more than 3,000 counties across the U.S., comparing lead concentrations in the air to homicide rates for the year 1990. Correcting for confounding social factors such as countywide income and education level, he and colleague Michael J. Lynch of the University of South Florida found that homicide rates in counties with the most extreme air-lead concentrations were four times as high as in counties with the least extreme levels (Arch. Pediatr. Adolesc. Med. 2001, DOI: 10.1001/archpedi.155.5.579).
Others have found similar correlations for U.S. cities, states, and even neighborhoods. In 2000, Rick Nevin, now a senior economist with ICF International, saw the trend for the entire country (Environ. Res., DOI: 10.1006/enrs.1999.4045). In general, these researchers see blood-lead levels and air-lead levels increase, peak in the early 1970s, and fall, making an inverted U-shape. About 18 to 23 years later, when babies born in the ’70s reach the average age of criminals, violent crime rates follow a similar trajectory.
Still, “predicting crime trends is hard,” Stretesky says. Anything that’s followed a U-shape over the same period is going to correlate, he says. One example put forward in a 2013 Mother Jones article titled “America’s Real Criminal Element: Lead” is vinyl record sales. They rose after World War II and then declined in the 1980s and ’90s, but that doesn’t mean they’re responsible for crime trends.
Seeking to strengthen the provocative lead-crime argument, in 2007, Nevin looked abroad, at countries where the crime rates didn’t necessarily follow the inverted U pattern. In every case—New Zealand, West Germany, Italy, the U.K., and so on—the data plots for blood-lead levels overlaid with plots of violent crime rates that were shifted back about 23 years (Environ. Res. 2007, DOI: 10.1016/j.envres.2007.02.008).
“When people read about my work,” Nevin says, “they oftentimes blurt out, ‘Correlation does not mean causation.’ ” One of the key signs of causation, he argues, is biological plausibility.
Research has shown that lead exposure does indeed make lab animals—rodents, monkeys, even cats—more prone to aggression. But establishing biological plausibility for the lead-crime argument hasn’t been as clear-cut for molecular-level studies of the brain. Lead wreaks a lot of havoc on the central nervous system. So pinpointing one—or even a few—molecular switches by which the heavy metal turns on aggression has been challenging.
What scientists do know is that element 82 does most of its damage to the brain by mimicking calcium. Inside the brain, calcium runs the show: It triggers nerve firing by helping to release neurotransmitters, and it activates proteins important for brain development, memory formation, and learning. By pushing calcium out of these roles, lead can muck up brain cell communication and growth.
On the cell communication side of things, lead appears to interfere with a bunch of the neurotransmitters and neurotransmitter receptors in our brains. One of the systems that keeps popping up in exposure experiments is the dopamine system. It controls reward and impulse behavior, a big factor in aggression. Another is the glutamate system, responsible in part for learning and memory.
On the brain development side of things, lead interferes with, among other things, the process of synaptic pruning. Nerve cells grow and connect, sometimes forming 40,000 new junctions per second, until a baby reaches about two years of age. After that, the brain begins to prune back the myriad connections, called synapses, to make them more efficient. Lead disrupts this cleanup effort, leaving behind excess, poorly functioning nerve cells.
“If you have a brain that’s miswired, especially in areas involved in what psychologists call the executive functions—judgment, impulse control, anticipation of consequences—of course you might display aggressive behavior,” says Kim N. Dietrich, director of epidemiology and biostatistics at the University of Cincinnati College of Medicine.
Dietrich and his colleagues have been studying lead’s effects on the developing brain for more than 30 years. In the late 1970s, he and a group of other investigators recruited some 300 pregnant women for what would become the Cincinnati Lead Study. At the time, these women lived in parts of Cincinnati—typically the inner city—that had experienced historically high numbers of lead-poisoning cases. Once the recruits’ babies were born, Dietrich and his group began monitoring the newborns too.
From the time they were born until they were six-and-a-half years old, the young participants had their blood-lead levels measured 23 times. The average childhood concentration for the whole group was 13 µg/dL. Now adults in their 30s, the subjects are having their brains scanned and behaviors analyzed.
And the results are eerie. As of 2008, 250 members of the lead study had been arrested a total of 800 times. The participants’ average blood-lead levels during childhood also correlated with their arrest rate, Dietrich’s team found (PLoS Med. 2008, DOI: 10.1371/journal.pmed.0050101).
Working with Dietrich, Kim M. Cecil, an imaging expert at Cincinnati Children’s Hospital Medical Center, has taken magnetic resonance images of the subjects’ brains and found that as childhood blood-lead levels increase, gray matter volume decreases in a handful of brain areas (PLoS Med. 2008, DOI: 10.1371/journal.pmed.0050112). Even more important, the regions with the largest gray matter loss are the ventrolateral prefrontal cortex and the anterior cingulate cortex, areas known for impulse control, emotional regulation, and decision making.
“These are the parts of the brain that say, ‘Ooh, I’ve learned from before that I shouldn’t steal that, or if I do this, then the consequences are that,’ ” Cecil says.
Still another way lead might coax the brain into committing violent acts is through IQ and learning disabilities. Although controversial when they were first reported, studies have shown that a child’s blood-lead level is inversely proportional to IQ. The extent of this relationship is still a point of contention, but most estimates have suggested that for every 10 µg/dL of blood lead, a child loses between one and 10 IQ points.
This might not seem like a lot for someone who’s been genetically gifted with an IQ around 120 or 110, Cecil contends. But for a child who might have started life with an IQ of 80, dropping to 70—a value close to impairment—is a handicap, she says.
Children who perform poorly in school and who have learning disabilities tend to have low self-esteem, get frustrated more easily, and, thus, are more likely to act out and engage in delinquent behavior, experts say.
On the molecular level, lead might be affecting learning and intellect through the N-methyl-
“Lead’s a potent inhibitor of NMDAR,” Guilarte says. That’s a problem, he says, because “NMDAR is crucial for brain development, learning, and memory processes.”
Over the years, Guilarte and his team have discovered that lead somehow decreases the number of NMDARs anchored to the surfaces of nerve cells in synapses. The heavy metal binds to and disables some of the receptors, too, preventing them from ushering calcium into the nerve cells on which they reside. This in turn prevents calcium from activating enzymes such as calmodulin kinase, a protein inside the cells that goes on to participate in strengthening nerve cell connections when a person learns a fact or commits an experience to memory (Brain Res. Rev. 2005, DOI: 10.1016/j.brainresrev.2005.02.004).
“Overall, the evidence is sufficient that early exposure to lead triggers a higher risk for engaging in aggressive behavior,” says U of Cincinnati’s Dietrich. “The question now is, what is the lowest level of exposure where we might see this behavior?”
Most kids in the U.S. today have a blood-lead level of 1 or 2 µg/dL. But there are nearly a half-million children between the ages of one and five with a blood-lead level above the 5-µg/dL threshold. These are mostly kids who are growing up in dilapidated inner-city houses with lead paint still on the walls or in neighborhoods with elevated levels of lead in the soil.
Despite progress in lowering lead levels in the environment, these kids would benefit from the reevaluation of crime policies and reinvigoration of cleanup efforts, says U of Colorado’s Stretesky. “People who are suffering the most from lead exposure are those that tend to be poor, minority, and low income.”