Montelukast, sold originally under the brand name Singulair, works differently from inhaled steroids. Rather than calming airway inflammation directly, it blocks leukotriene receptors, molecules involved in the inflammatory cascade that narrows airways during asthma attacks. Because it comes as a once-daily pill rather than an inhaler, it is far easier to give to young children, and adherence rates reflect that convenience. In prescription claims data, the median proportion of days covered among children who filled at least one prescription was 52 percent for montelukast but only 27 percent for ICS. That practical advantage has made the drug an increasingly popular second-line controller worldwide, though prescribing patterns vary dramatically: in the United States roughly 10.5 to 16.5 percent of children with asthma use montelukast, while in South Korea the figure reaches 44.8 percent.

The drug’s neuropsychiatric reputation, however, has been contested for years. Spontaneous adverse event reports have linked montelukast to sleep disturbances, hyperactivity, attention problems, depression, anxiety, agitation, aggression, irritability, stuttering, seizures, and even suicidal behavior. The FDA responded in 2020 with a boxed warning about serious mental health side effects. Yet the large observational studies that followed produced contradictory results. A US cohort study by Paljarvi and colleagues found that adding montelukast to ICS treatment raised the incidence of neuropsychiatric events by about a third, while a Swedish study by Wintzell and colleagues, comparing montelukast against long-acting beta-agonists, found no elevated risk at all. The new research was designed specifically to resolve why such studies disagree.

The research team, led by Xue-Feng Jiao and senior author Anick Bérard of CHU Sainte-Justine in Montreal, exploited a methodological advantage that only birth cohorts can offer: complete health records stretching back to each child’s birth. The American cohort, AM-PREGNANT, drew on the MarketScan Commercial Database covering more than 2.5 million children born between 2003 and 2021, while the Quebec Pregnancy Cohort followed children born between 1998 and 2015. From these populations the investigators identified children aged 1 to under 18 whose first-ever asthma diagnosis was followed by either montelukast monotherapy or ICS monotherapy. Crucially, they excluded any child with a pre-existing psychiatric or neurological diagnosis, prior psychotropic medication use, or chronic diseases such as diabetes or congenital heart disease, eliminating confounders that plagued earlier work.

The numbers were striking. In the American cohort, 85,322 asthmatic children were included, with 21,534 starting on montelukast and 63,788 on ICS. The Quebec cohort contributed 11,854 children, of whom 373 received montelukast. The median age at treatment start was around two and a half years in both cohorts, a detail that matters because asthma most often begins before age six, and the neuropsychiatric outcomes examined, including sleep disorders and ADHD, typically emerge in early childhood. Children were followed until a first diagnosis of an outcome, a change in controller medication, discontinuation of treatment, or the end of their insurance enrollment.

When the two cohorts were pooled through random-effects meta-analysis, montelukast initiation was associated with a 40 percent higher risk of ADHD compared with ICS, with a pooled adjusted hazard ratio of 1.40 and a 95 percent confidence interval of 1.19 to 1.66. The risk of oppositional defiant and conduct disorders, conditions marked by defiance, temper tantrums, aggression, and destructiveness, was 43 percent higher, with a hazard ratio of 1.43. By contrast, the analysis found no significant differences between the drugs for mood disorders or anxiety disorders. The median time from starting montelukast to an ADHD diagnosis was 640 days in the American cohort but only 131 days in Quebec, and sleep disorder diagnoses in Quebec appeared a median of just 21 days after the first prescription fill.

Sleep disorders told a more complicated story. The two cohorts disagreed sharply, with substantial statistical heterogeneity between them: Quebec showed a dramatically elevated risk, with a hazard ratio of 5.51, while the American cohort showed essentially no association. The researchers traced the discrepancy to differences in asthma severity. Over 90 percent of Quebec children had controlled or partly controlled asthma, compared with only 44 percent in the American cohort, and only 2 percent of Quebec children used oral corticosteroids during follow-up versus roughly half in the United States. When the team stratified children by asthma control, montelukast was linked to increased sleep disorder risk precisely in the subgroup with controlled or partly controlled asthma, and among children who never needed oral corticosteroids. Severe, poorly controlled asthma itself disrupts sleep, the authors explain, which can mask any additional drug effect.

Seven sensitivity analyses tested the robustness of the findings, and the core results held. The associations persisted when children with any prior montelukast or ICS exposure were excluded, when outcome definitions were broadened to include prescriptions for hypnotic or ADHD medications, when grace periods for medication adherence were redefined, when children with less than 30 days of cumulative treatment were removed, when the asthma definition was loosened, and when propensity score weighting and matching replaced conventional regression. The researchers also applied validated correction factors to adjust the recorded duration of ICS prescriptions, a known weakness of claims data, and used the validated Pharmacoepidemiologic Pediatric Asthma Control Index to account for asthma severity, addressing the confounding-by-indication problem that has undermined previous studies.

Why might a leukotriene receptor antagonist influence the developing brain? The mechanism remains unresolved, but several biological hypotheses exist. Network pharmacology analyses have identified dozens of intersecting molecular targets of montelukast associated with sleep disorders, implicating pathways ranging from receptor activation in chemical carcinogenesis to estrogen signaling. Other proposed mechanisms include interference with the glutathione detoxification system in the brain, dysregulation of neurotransmitter and neurosteroid pathways involving the hypothalamic-pituitary-adrenal axis, and disruption of mitochondrial function in neurons. The overlap between ADHD and disruptive behavior disorders, which co-occur in roughly half of children with ADHD, suggests that montelukast-induced ADHD and montelukast-induced oppositional and conduct disorders may share common underlying mechanisms, consistent with the hostility, irritability, and restlessness already listed in the FDA-approved drug labeling.

The authors caution that their study cannot fully exclude residual confounding, that prescription fills may not equal actual intake, and that the mostly preschool-aged population prevented assessment of later-emerging outcomes such as psychosis or self-harm. Still, the design, capturing children at the very first asthma diagnosis with longitudinal data from birth, offers some of the cleanest evidence yet on this safety question. The practical message for clinicians and families is vigilance rather than alarm: when montelukast is prescribed, parents and physicians should watch for insomnia, dream abnormalities, sleepwalking, inattention, hyperactivity, impulsivity, temper tantrums, defiance, hostility, and aggression, and weigh these risks against the drug’s genuine convenience advantage, particularly for children whose asthma is already well controlled.

Subject of Research: Neuropsychiatric safety of montelukast compared with inhaled corticosteroids in children with newly diagnosed asthma

Article Title: Neuropsychiatric outcomes after montelukast versus inhaled corticosteroids in pediatric asthma: a pooled birth cohort study

Article References: Jiao, X.-F., Ducharme, F. M., Leal, L. F., Longo, C., Sheehy, O., Kaguelidou, F., & Bérard, A. (2026). Neuropsychiatric outcomes after montelukast versus inhaled corticosteroids in pediatric asthma: a pooled birth cohort study. BMC Medicine, 24(1), Article 536. https://doi.org/10.1186/s12916-026-05235-7

Image Credits: AI Generated

DOI: 10.1186/s12916-026-05235-7

Keywords: montelukast, pediatric asthma, ADHD, sleep disorders, oppositional defiant disorder, conduct disorder, inhaled corticosteroids, drug safety, pharmacoepidemiology, birth cohort, FDA boxed warning, neuropsychiatric adverse events