The documented risk of untreated fever far exceeds any theoretical risk to neurodevelopment.
For years, a pregnant woman reaching for acetaminophen has done so against a backdrop of doubt sown by alarming headlines and contested science. A major 2026 review in Obstetrics & Gynecology Science now restores clarity: examining data from millions of children across Sweden and Japan, researchers found no meaningful link between prenatal acetaminophen use and autism or ADHD once family and genetic factors were properly accounted for. The finding does not merely exonerate a medication — it reminds us that fear, when it displaces evidence, can itself become a source of harm.
- Media coverage and political rhetoric had turned a routine pregnancy decision into a source of genuine clinical anxiety for both patients and physicians.
- Earlier studies claiming two-to-three times higher neurodevelopmental risk were built on flawed methods — measuring drug exposure only at delivery, not across the full arc of pregnancy.
- A sibling-comparison analysis of 2.48 million Swedish children produced hazard ratios of 0.98 for both autism and ADHD — numbers statistically indistinguishable from zero risk.
- Untreated maternal fever carries documented dangers including birth defects and miscarriage, making the real clinical risk the avoidance of the medication, not its use.
- WHO and ACOG guidelines hold firm: acetaminophen remains the first-line choice, to be used at the lowest effective dose for the shortest necessary time under medical supervision.
A pregnant woman with a fever faces a choice that should be simple — but in recent years, headlines and social media have cast doubt on acetaminophen, long considered the safest painkiller in pregnancy, by suggesting a link to autism in children. That cloud of uncertainty has left patients and clinicians second-guessing a medication with decades of established safety.
A review accepted in May 2026 in Obstetrics & Gynecology Science examines what the evidence actually shows. The authors focused on population-based studies using sibling-comparison designs — an approach that controls for the shared genetic and environmental factors that distort ordinary observational research. Their primary source was a Swedish cohort of more than 2.48 million children tracked over decades, supplemented by Japanese data covering 200,000 participants, with outcomes including autism spectrum disorder, ADHD, and intellectual disabilities.
The conclusion is unambiguous: once family context is accounted for, the apparent risk vanishes. The Swedish study returned hazard ratios of 0.98 for both autism and ADHD — effectively no increased risk. Earlier studies that had reported two-to-three times higher rates were found to be methodologically compromised, relying on cord blood samples that captured only delivery-time exposure and failing to control for differences between mothers who use fever-reducing medication and those who do not.
The broader picture matters. Rising autism prevalence reflects expanded diagnostic criteria and greater awareness, not painkiller use. Untreated maternal fever poses real, documented dangers — birth defects, miscarriage, premature delivery — that dwarf any theoretical neurodevelopmental concern. Acetaminophen has long been the first-line choice precisely because it is safer than the alternatives. The guidance remains: lowest effective dose, shortest necessary duration, always under medical supervision. In a climate of health anxiety and information overload, this review is a case study in why good medicine demands reading past the headline.
A pregnant woman with a fever faces a choice that should be straightforward but has become clouded by headlines and social media alarm. Should she take acetaminophen—the painkiller sold as Tylenol—or endure the fever? In recent years, media coverage and political statements have suggested a direct link between the drug and autism in children, creating a wave of clinical uncertainty that has left both patients and doctors second-guessing a medication that has long been considered the safest option during pregnancy.
A new review, accepted for publication in May 2026 in Obstetrics & Gynecology Science, cuts through that noise by examining what the actual evidence shows. The work is a narrative review that looked critically at population-based studies, with particular attention to research using sibling-comparison designs—a methodological approach that controls for shared genetic and environmental factors that often confound ordinary observational studies. The authors examined data from a Swedish cohort of more than 2.48 million children tracked over decades, along with Japanese studies involving 200,000 participants. They monitored diagnoses of autism spectrum disorder, attention-deficit/hyperactivity disorder, and intellectual disabilities.
The findings are clear: when researchers adjusted their analysis to account for family context, the supposed risk disappeared. In the major Swedish study, the hazard ratios—a measure of risk similar to odds ratios—were 0.98 for autism and 0.98 for ADHD, with confidence intervals that included 1.0. Numbers this close to 1.0 mean there is no statistically significant increase in risk from the medication. Earlier studies that had suggested risks two to three times higher were found to have serious methodological flaws. Some relied on cord blood biomarkers that only captured drug exposure at the moment of delivery, not throughout the entire pregnancy. Others failed to account for the fact that mothers who use fever-reducing medication might differ in important ways from those who don't.
The broader context matters here. Autism prevalence has risen globally, but this reflects wider diagnostic criteria and greater awareness, not increased use of painkillers. Managing pain and fever during pregnancy is not a luxury—it is essential care. Untreated maternal fever can lead to serious complications including birth defects, miscarriage, and premature delivery. For decades, acetaminophen has been the first-line choice precisely because its safety profile is superior to nonsteroidal anti-inflammatory drugs and opioids.
The review echoes recommendations from the World Health Organization and the American College of Obstetricians and Gynecologists. The practical message is simple: acetaminophen remains the safest option. When a pregnant woman has a fever, the documented risk of leaving it untreated far exceeds any theoretical risk to her child's neurodevelopment. The guidance should be equally clear: use the lowest effective dose for the shortest necessary time, respecting a maximum of 4,000 milligrams daily, always under medical supervision. In an era of information overload and health anxiety, this review serves as a reminder that good medicine requires reading past the headlines to understand what the evidence actually says.
Notable Quotes
The risk of not treating a high maternal fever is comprovably superior to any theoretical risk of neurodevelopmental harm.— Review findings, echoing WHO and ACOG recommendations
The Hearth Conversation Another angle on the story
Why did this review become necessary? Wasn't acetaminophen already considered safe in pregnancy?
It was, for decades. But in recent years, media coverage and political statements created a narrative linking the drug to autism. Patients started refusing it, doctors started hesitating. The science hadn't changed—the noise around it had.
What makes the sibling-comparison studies different from the earlier ones that suggested higher risks?
The earlier studies compared pregnant women who used acetaminophen to those who didn't, but those groups differ in countless ways—income, education, health status, reasons for having fever. Sibling studies compare children in the same family, one exposed to the drug in utero and one not. That controls for almost everything genetic and environmental that's shared.
So the hazard ratios of 0.98 mean what, exactly?
They mean no increased risk. A ratio of 1.0 means no difference between groups. These numbers are so close to 1.0, and their confidence intervals cross 1.0, that we can't say the drug causes harm. The earlier studies showing two to three times higher risk were measuring something else—probably confounding factors, not the drug itself.
What was wrong with using cord blood biomarkers?
Cord blood only shows what's in the baby's system at birth. If a mother took acetaminophen in the first trimester, it won't show up in cord blood. You're only capturing exposure at delivery, not the full pregnancy picture. It's like trying to understand someone's diet by looking at what they ate on one day.
Why does untreated fever matter so much?
High maternal fever is documented to cause birth defects, miscarriage, and premature delivery. Those are real, proven harms. The theoretical risk from acetaminophen—which this review shows doesn't actually exist—is far smaller than the known risk of doing nothing.
What should a pregnant woman do if she has a fever tomorrow?
Take acetaminophen under her doctor's guidance. Use the lowest dose that works, for as short a time as needed, up to 4,000 milligrams a day. The fear that's been created around this drug is the real problem, not the drug itself.