How High Is Too High? Dosing ADHD Medications

In children, the benefits rise with the dose, but only up to a point (dashed line = FDA-approved max).

Most guidelines tell us to titrate up but say little about when to stop.

STUDY: Nourredine M et al, Lancet Psychiatry 2026;13:485–495

STUDY TYPE: Dose-effect network meta-analysis

FUNDING: National Institute for Health and Care Research

Background

Does ADHD respond better as the dose goes up? This analysis looks at dose-response curves to see where benefits peaks and tolerability starts to slip. It expands on a similar study of adults from 2024 by including non-stimulants and children.

The Analysis

• 113 randomized controlled trials: 68 in children and adolescents (14,138 participants) and 45 in adults (11,016 participants)
• Stimulants (methylphenidate, amphetamines) and non-stimulants (guanfacine, atomoxetine, modafinil, viloxazine) evaluated across fixed and flexible dose ranges
• Ampetamines are given as dextroamphetamine equivalents

In children and adolescents, methylphenidate peaked at about 45 mg/day, amphetamines at 25 mg/day, and guanfacine at 4 mg/day, with no added benefit at higher doses. For amphetamines in children, discontinuation risk climbed above 25 mg/day, the same point where benefit plateaued.

In adults, amphetamines plateaued around 50 mg/day (FDA maximum: 40 mg/day). Methylphenidate in adults showed continued gains up to 50 mg/day but with sharply rising discontinuation risk above that threshold: 7.3% at 60 mg/day versus 2.6% for placebo.

Lisdexamfetamine (Vyvanse)

This prodrug converts into dextroamphetamine at a steady rate, creating smooth levels throughout the day that avoid the up-and-downs of other formulations. In theory, you can get more out of a smaller daily dose, because the dose isn’t wasted in those peak levels. That is why the FDA’s upper limit (70 mg Vyvanse) is equivalent to a relatively low dextroamphetamine dose (28 mg dextroamphetamine), well below that drug’s FDA-max of 40 mg.

The analysis didn’t address this, as they weren’t able to establish a clear dose-response curve for lisdexamfetamine (Vyvanse).

Non-Stimulants

Guanfacine showed increasing efficacy up to 4 mg/day in children and adolescents. There wasn’t enough data for this one in adults, nor for clonidine.

Atomoxetine and modafinil didn’t have a clear dose-effect relationship.

Comparison to Earlier Studies

The results are similar to two analyses by Farah et al, which also found upper limits in children/adolescents and adults.

Key difference is that Farah found no extra benefit beyond 30-35 mg amphetamines (as dextroamphetamine dosage) in adults, while this study found the benefits ceased around 50 mg. That sounds like a big difference, but the dose-response looks pretty similar in the two graphs. Also, both papers show higher drop-out rates above 35 mg amphetamines, which is also the cut-off where we see more mania and psychosis.

In children and adolescents, Farah et al found similar upper limits for stimulants, where the benefits plateaued beyond 30 mg of methylphenidate (vs 45 mg in current study) or 20 mg of amphetamines (vs 25 mg in current study).

For atomoxetine, earlier studies found a peak benefit at 1.2 mg/kg per day; for viloxazine, 400 mg/day (both in children and adolescents).

Limitations

The adult dataset was small. Analyses were short-term (7–8 weeks on average). Long-term cardiovascular risks at high doses were not captured.

Practice Implications

• All analyses find a maximum upper limit for benefits on stimulants.
• These upper limits are averages, and have to be personalized for the patient. But if all your patients are dosed above average, time to rethink things.
• One randomized trial found higher doses of amphetamines were not effective, unless they had severe ADHD.

Graphs

Note the dashed vertical line in these graphs represents the FDA-max, not the peak level from this analysis.

Dropouts by Dose: Children