Montelukast linked to 40% higher ADHD risk, versus inhaled steroids, in 97,176 children with new asthma

For a child with asthma, montelukast has an obvious appeal. It is a pill or chewable tablet taken once a day, and in a household wrestling a toddler and an inhaler, montelukast became an increasingly popular second-line controller, the daily preventer doctors reach for when the usual first choice, an inhaled steroid, does not suit. In the US cohort of the new study, a quarter of children with new asthma were started on it first. It also carries the strongest warning the US Food and Drug Administration puts on a drug. Since 2020 its label has said that serious mental health side effects, including suicidal thoughts or actions, have been reported in people taking it.

What has been missing is a clear answer on how often such problems follow its use in children, because large studies have pointed in different directions. A study published on October 1 in BMC Medicine adds a sizable piece of evidence from the youngest patients. Researchers at the Sainte-Justine university hospital research center in Montreal, led by Xue-Feng Jiao with Anick Bérard as senior author, followed 97,176 children in the US and Quebec from their first asthma diagnosis. Children whose first daily asthma medicine was montelukast were about 40% more likely to be diagnosed with attention-deficit/hyperactivity disorder, or ADHD, than children started on an inhaled steroid. They were similarly more likely to be diagnosed with oppositional defiant or conduct disorder. Mood and anxiety disorders showed no difference. The study is observational, and two earlier large studies found no link, so it adds a signal without settling the question.

Why montelukast carries a boxed warning

Montelukast, sold as Singulair and as generics, blocks substances in the body called leukotrienes, which help cause the symptoms of asthma and allergies. The FDA approved it in 1998 and, after years of case reports, strengthened its warnings in March 2020. The agency listed a long run of possible behavior and mood changes, including aggression, attention problems, bad or vivid dreams, trouble sleeping, irritability and depression. For hay fever, it advised reserving the drug for people who cannot use other medicines.

The FDA’s own evidence was mixed. Its review drew mainly on reports sent in by doctors and families. The agency also noted that its Sentinel study and other observational studies did not find an increased risk compared with inhaled corticosteroids, while adding that those studies had limitations. Prescribing nonetheless shifted. An analysis of US insurance claims found that new montelukast prescriptions for asthma fell after the boxed warning, in every age group.

The same Montreal group had looked at the question once before, on a small scale. In 2017 a Sainte-Justine team interviewed the parents of 106 children who had started montelukast and found that 16% had stopped it because of neuropsychiatric reactions, mostly irritability, aggressiveness and sleep disturbances, usually within two weeks.

How the Montreal team followed children from their first asthma diagnosis

Earlier large studies, the authors argue, all had a weak point. Some compared children taking montelukast with an inhaled steroid against children on the steroid alone, which mixes in children with more severe asthma. Some enrolled children around age 9 or 12, by which time many had already tried several asthma drugs. Most could not cleanly exclude children who already had a psychiatric diagnosis.

To get around that, the team used two birth cohorts that track children’s health records from birth. One, called AM-PREGNANT, draws on US commercial insurance claims from 2003 to 2021; the other, the Quebec Pregnancy Cohort, uses Quebec’s public health insurance records from 1998 to 2015. In each, the researchers found children aged 1 to 17 with a first-ever, doctor-confirmed asthma diagnosis whose first controller medicine, the daily drug meant to prevent attacks, was either montelukast on its own or an inhaled corticosteroid on its own. Children with any earlier psychiatric or neurological diagnosis, or a month or more of psychiatric medicines, were left out.

The two groups were very different in size. In the US cohort 21,534 children started on montelukast and 63,788 on inhaled steroids; in Quebec, where steroid inhalers dominate, the numbers were 373 and 11,481. Most children were toddlers, with a median age of about two and a half at their first prescription. The analysis took account of a long list of differences. They included age, sex, allergies, obesity, prematurity and how well controlled the asthma was, along with the mother’s own psychiatric history, her smoking, drinking or drug use in pregnancy, and the child’s use of other medicines linked to behavior changes. Each cohort was analyzed separately, and the two results were pooled whenever a standard statistical test showed they were consistent enough to combine.

How much more often ADHD and conduct problems followed montelukast

Pooling the two cohorts, montelukast was associated with increased risks of ADHD. Children started on it had about 40% higher risk of an ADHD diagnosis than those started on an inhaled steroid. Oppositional defiant and conduct disorders, the diagnoses given for persistent defiance and temper outbursts and, at the more serious end, aggression and destructive behavior, were about 43% more likely. That second result comes from the US cohort alone, because no child on montelukast in Quebec received the diagnosis. For mood disorders and anxiety disorders there was no clear difference between the drugs.

Sleep was the messy result. In Quebec, sleep disorders were several times more common with montelukast, an estimate built on very few children and with a range of uncertainty stretching from slightly higher to more than 20 times higher. In the far larger US cohort the difference was small and could have been chance. When the team split children by how well controlled their asthma was, a higher risk of sleep disorders appeared in both cohorts among children whose asthma was controlled or partly controlled. The authors suggest that poorly controlled asthma disrupts sleep on its own, which could hide any extra effect of the drug in sicker children.

The results held up across seven extra checks, including one restricted to children who had never had either drug before their asthma diagnosis and others using two statistical methods for balancing the groups. The ADHD diagnoses came late. Among US children who were diagnosed, the median time from first montelukast prescription to diagnosis was 640 days, close to two years, against 21 days for the sleep problems seen in Quebec. Because children left the analysis when they stopped filling prescriptions, most were followed for only a few months, and the ADHD diagnoses came from the minority who stayed on the drug for a long time. The pattern fits a slowly emerging problem, or one that simply takes time to be recognized and diagnosed, more than an immediate reaction to the drug.

Why studies of montelukast and ADHD keep disagreeing

The new result does not stand alone, and it does not sit easily with its neighbors. In Taiwan, a national study that matched 12,806 children on montelukast with 12,806 similar children not taking it found almost identical ADHD rates, 4.94% and 4.76%, and concluded that montelukast does not increase the risk of the condition. In Sweden, a nationwide study compared 74,291 children aged 6 to 17 taking montelukast or another add-on asthma drug. It found “no association between use of montelukast and the risk of neuropsychiatric adverse events,” though its estimate for disrupted attention and behavior leaned slightly higher and was uncertain. A 2026 meta-analysis of 21 studies found no clear overall increase in neuropsychiatric risk in children, while it did find one in adults.

The Montreal team argues that its design explains the gap: younger children, followed from the very start of treatment, compared with the usual alternative first drug. The Swedish and Taiwanese researchers made different design choices, and each set of choices has its own blind spots. The Swedish study followed children for an average of under six months, which the Montreal authors say is short for conditions that take months or years to surface; their own median follow-up was shorter still.

There is also the question of who gets montelukast in the first place. Families who prefer a once-daily pill, children with allergies, and doctors who choose it for particular patients may all differ in ways a claims database cannot record, which is the problem epidemiologists call confounding by indication. The authors adjusted for allergy and asthma control and found the ADHD link in children with and without allergies, but they acknowledge they cannot completely rule out residual confounding, meaning other differences between the groups.

What the montelukast cohort study leaves out

The follow-up was short. Because children were counted only while they kept filling their prescription, the median follow-up was 120 days in the US cohort and 30 days in Quebec. A diagnosis of ADHD in a toddler is itself uncommon and often comes later. The analysis relied on diagnosis codes in insurance records, which miss milder cases, and on pharmacy fills, which do not show whether a child took the medicine. Most children were under 6, so outcomes that appear later, such as suicidal thoughts, could not be studied.

The headline figures are also relative. They say how much more often ADHD was diagnosed in children on montelukast, not how many children that amounts to, which depends on how common the diagnosis is at that age.

What this means for the montelukast warning

The study does not change the FDA’s warning, which already lists attention problems among the side effects to watch for. The authors suggest that when montelukast is prescribed, “clinicians and families may need to monitor children for symptoms of sleep disorders, ADHD, and oppositional defiant and conduct disorders.” They also say their results “need to be replicated in other populations.”

For families, the choice between a once-daily pill and an inhaled steroid involves how well each controls a particular child’s asthma, how reliably it gets taken, and side effects on both sides, and it is assessed by the child’s doctor. Do not stop or change a prescribed medicine without talking to your doctor.

The new data add a signal to the warning on montelukast in young children, while leaving open whether the drug itself is the cause.