Jarlenski’s research has also shown women’s perception of cannabis as risky is dropping. And a study published in June in the Journal of the American Medical Association (JAMA) found that between 2002 to 2003 and 2016 to 2017, self-reported use of cannabis in pregnancy doubled overall in the U.S., from 3.4 percent to 7 percent. Similarly, a 2017 study of nearly 30,000 Kaiser patients in Northern California found cannabis use during pregnancy, either self-reported or measured in urine toxicology tests, increased from 4.2 percent in 2009 to 7.1 percent in 2016. Urban medical centers have found as many as 22 to 27 percent of pregnant women are positive for cannabis.

Other cannabis researchers report concerns about Dreher’s study, particularly if it’s being used to inform decisions today. “It’s a very small sample size. There’s definitely been studies that have been done with many more children that have looked at similar outcomes that haven’t had the same findings,” Metz said. In the same June issue of JAMA, an Ontario study analyzed the medical records of more than 660,000 pregnancies and found those who reported using cannabis were also more likely to have preterm birth, placental abruption, and small for gestational age babies, with lower Apgar scores — which indicate general health based on factors including muscle tone and breathing — along with greater admission to the neonatal intensive care unit (NICU). 
 
“The ‘Jamaica study’ continues to garner fame despite being one small study in a relatively large pool of evidence,” said Metz.

And when the National Academies of Sciences, Engineering, and Medicine summarized the health effects of cannabis in a 2017 report, they concluded there’s substantial evidence for an association with lowered birth weight and limited evidence for an association with pregnancy complications for the mother and infants needing NICU care. For a range of other outcomes, they said that the evidence was insufficient to make conclusions, because it was conflicting or had too many methodological limitations.

Another issue with Dreher’s study, and any study of its age, is that cannabis potency has increased dramatically over time. “I would throw those studies out completely. I don’t think you could go back to those early studies and say they have many implications for what is going on today,” said Barry Lester, a professor of psychiatry and pediatrics and director of the Center for the Study of Children at Risk at Brown University. In the U.S., the THC content of illicit cannabis samples increased from 4 percent in 1995 to 12 percent in 2014, and legal cannabis obtained in 2017 in Washington state was around 20 percent THC, with some products potentially reaching 30 percent or higher. There’s little data on how this might impact fetal exposure, but a 2015 abstract described an increase in the concentration of THCA (a THC metabolite) measured in the first bowel movements of Colorado newborns from 213 nanograms per gram (ng/g) in 2012 to 361 ng/g in 2014.

The cannabis used by the women in Dreher’s study wasn’t high in THC, she said. “It was not anything like the cannabis that we have now.”

Lester also collaborated with Dreher in Jamaica in the 1980s, studying the same cohort of babies from her more well-known paper. His study, published with Dreher as a coauthor in Child Development in 1989, found that at four and five days of age, the cannabis-exposed newborns had shorter, higher-pitched cries with more dysphonation and variability. These were the same babies with no difference in behavior at three days old, but Lester said that didn’t surprise him, because the behavioral scales used aren’t typically designed to be as sensitive. “Cry acoustics has a long-standing track record of differentiating at-risk kids,” he said. “If you pick a dependent measure that you know has that kind of sensitivity, you’re obviously more likely to find effects.”

Yet for all their criticism of Dreher’s study, these researchers also wanted to talk about the limitations of the current science and the need for nuanced interpretation.

Because it would be unethical to randomly assign women to use cannabis while pregnant, the evidence we have on this question is observational and rife with confounding factors. In North America, those who use cannabis are also more likely to use tobacco, alcohol, and other drugs, and to face challenges that come with poverty, mental illness, and effects of racism. Researchers try to statistically account for these differences (indeed, each of these factors was identified and accounted for in the Ontario JAMA study), but quantifying all the inequities impacting mothers and their children is a tall order, and that’s a real limitation of the evidence we have.

That adds uncertainty to our understanding of the health risks of cannabis, said Jarlenski, the University of Pittsburgh researcher. Still, she rattled off a host of other reasons not to use cannabis while pregnant, including unstandardized potency, lack of evidence of benefits, risk of dependency, and unknowns of how it might interact with postpartum depression. There’s also a risk of being reported to child welfare services for using a substance that’s still considered a Schedule I drug at the federal level, and poor women of color are likely at greatest risk of this consequence.

“I would throw those studies out completely,” says Lester, because cannabis potency has increased dramatically over time.

At the same time, Jarlenski cautions that it’s important not to inflate the possible risks of prenatal cannabis use. This occurred with the so-called “‘crack baby’ epidemic” of the 1980s, she said, when the effects of prenatal cocaine use were exaggerated and ultimately did more harm than good to women and their families. Pregnant women need to know about possible risks of cannabis, but some may also need public health resources to help them quit, she added. Others may need advice about safer strategies for managing pregnancy symptoms, Metz said.

The largest longitudinal study of cannabis use in pregnancy includes nearly 10,000 pregnant women in Rotterdam, the Netherlands, where cannabis is available to purchase in coffee shops and is high in THC. That study found that cannabis use in pregnancy was associated with decreased fetal growth, above and beyond the effects of tobacco. At ages 7 to 10, the same kids were more likely to have externalizing problems, like aggressive behavior and rule-breaking, and psychotic-like experiences, like hearing voices. On MRI brain scans, they had thicker prefrontal cortices.

However, Henning Tiemeier, a principal investigator for the study and a professor at the Harvard T.H. Chan School of Public Health, is cautious in interpreting those results. With careful analysis, he and colleagues have concluded that the behavioral and brain differences in cannabis-exposed kids are probably not caused by cannabis but are more likely related to genetic or environmental factors, because they see the same associations in kids whose mothers used cannabis before but not during pregnancy, and in kids whose fathers used cannabis. The decrease in fetal growth couldn’t be explained by these other factors, though; it seems independently associated with in-utero exposure to cannabis, perhaps because it impacts uterine blood flow. This finding is one of the most consistent across studies of cannabis use in pregnancy, especially among heavier users.

Tiemeier emphasizes that though their work is somewhat reassuring in terms of long-term effects on kids, the question is far from resolved. Answering it has been impeded by a lack of funding and barriers to studying cannabis. Meanwhile, he says, discussions of cannabis are being driven by people who cherry-pick the science that supports what they want to believe. “I’m surprised how many people are opinionated on the basis of very poor facts,” he said.

“We are saying this again and again: Those that say cannabis is safe during pregnancy do not really know. Those that say it is not [safe] do not know either,” Tiemeier said. It’s a scientific debate that can’t be resolved with the results of one study, be it a tiny one in rural Jamaica or a large one in a modern Dutch city.