Dosage of Non-Stimulant Medications for ADHD

Completely revised 09/2026

 

Non-stimulant medications such as atomoxetine and guanfacine may be considered if stimulants are ineffective, not well tolerated, or if all-day efficacy is needed.

The most important difference: It takes several weeks for non-stimulant medications to take full effect. They must also be taken consistently, not just as needed, and tapered off gradually rather than stopped abruptly.

Non-stimulants are the third-line treatment for ADHD, after the stimulants AMP and MPH.

Clinicians report that non-stimulants sometimes lose their effectiveness after 9 to 18 months, at which point it becomes necessary to switch to another non-stimulant. (E 4)¹ This is explicitly a clinical observation. Published research consists almost exclusively of short-term studies on efficacy.

With regard to atomoxetine, the available data tend to suggest that there is no loss of efficacy. A review of response patterns notes that the effect continues to build over a period of up to 24 weeks and beyond, and that relapse rates after discontinuation are lower than expected: In adults, 50% maintained their response for at least six months after discontinuation if they had previously been treated for six months. A single dose can produce a 24-hour effect, even though the plasma half-life is only about five hours; thus, the effect does not correlate with plasma levels. The authors attribute this to neuroadaptive changes.² In the two 10-week approval studies in adults, approximately 50% responded to atomoxetine (at least a 25% reduction in the CAARS total score); in children and adolescents across six studies, approximately 60% responded (at least a 25% reduction in the ADHD-RS). When it comes to discontinuation, atomoxetine differs significantly from stimulants. After nine months of atomoxetine followed by a switch to placebo, only 12.2% experienced a relapse, compared to 2.5% among those who continued treatment. In adults, 50.0% of those switched to placebo maintained their response for 25 weeks. In contrast, among those treated with lisdexamfetamine, 67.5% of children and 75% of adults experienced a relapse after switching to placebo, predominantly within two weeks. A tapering trial proceeds fundamentally differently with atomoxetine than with stimulants. The absence of a relapse while on atomoxetine therefore does not prove that the treatment was unnecessary. In a six-week crossover study, 22% did not respond to either OROS methylphenidate or atomoxetine. Of those who did not respond to OROS methylphenidate, 43% responded to atomoxetine; conversely, 42% of those who did not respond to atomoxetine had previously responded to OROS methylphenidate.

1. Dosage of Atomoxetine¶

Atomoxetine is available as non-divisible capsules and as divisible tablets. The capsules should not be opened, as the active ingredient in powder form can irritate the eyes.

Unlike methylphenidate, it is important to take atomoxetine consistently.(E 4)³

1.1. Starting dose for atomoxetine¶

(E 1b): According to the available studies (E 4)⁴ as well as our own experience, gradual up-dosing also offers the advantage of reducing the incidence of side effects with atomoxetine. While a manufacturer-sponsored study does not provide a clear picture (E 2b)⁵, other studies report that discontinuation due to side effects is less common with slow dose titration, even though the incidence of side effects was comparable. (E 1b)⁶

(E 4): In some cases, it is recommended to start with a dose of 10 mg for women or 18 mg for men—increasing the dose every 14 days (E 4)³ —rather than 40 mg/day, or to begin with an initial dose of 18 to 25 mg.(E 4)⁷

1.1.1. Slow Metabolizers and Atomoxetine: A Serum Level Test?¶

In a meta-analysis of 22 pediatric and 3 adult studies, side effects occurred earlier with once-daily dosing than with twice-daily dosing, and earlier with rapid up-dosing than with gradual up-dosing. In children, these side effects included abdominal pain, loss of appetite, and fatigue; in adults, they included insomnia, which also lasted longer. Loss of appetite and nausea lasted longer with once-daily dosing. Therefore, to avoid side effects, it is better to split the dose into two administrations and increase the dose more slowly.⁸

A large portion of the variability in atomoxetine response is genetically determined. Slow metabolizers of CYP2D6 (approximately 7% of the population of European descent) achieve approximately ten times higher total exposure and five times higher peak concentrations at the same dose. The half-life increases from about 5 to about 24 hours. Potent CYP2D6 inhibitors such as paroxetine, fluoxetine, or quinidine cause the same shift. The FDA label recommends that for known slow metabolizers, as well as when such inhibitors are co-administered, treatment should begin at 0.5 mg/kg/day and the dose should not be increased to the target dose of 1.2 mg/kg/day until after four weeks, that is, to explicitly slow the up-dosing rather than reduce the target dose.⁹

Whether a 50% dose reduction is sufficient for poor metabolizers is a matter of debate. An analysis of therapeutic drug monitoring data found an almost tenfold difference in the ratio of serum concentration to dose between poor and normal CYP2D6 metabolizers. The authors therefore consider the 50% reduction recommended by the Clinical Pharmacogenetics Implementation Consortium to be significantly too low. In addition, carriers of the CYP2C19*2 allele had, on average, 50% higher serum atomoxetine levels, regardless of their CYP2D6 genotype.¹⁰

The international pharmacogenetic guideline for atomoxetine does not recommend a blanket adjustment to the initial dosing regimen for poor metabolizers, but rather a plasma-monitoring-based approach: If no response is observed after two weeks, the plasma concentration should be measured and used, together with the CYP2D6 genotype, to determine the appropriate dose. A therapeutic range of 200 to 1,000 ng/ml is recommended. If the peak concentration is below 200 ng/ml, the dose should be increased proportionally toward 400 ng/ml. Above 400 ng/mL, a moderate further improvement was observed. Doses exceeding 120 mg/day have not been extensively studied but may be necessary to achieve the target concentration.¹¹

The Dutch pharmacogenetic working group has issued a different recommendation. In the absence of CYP2D6 activity, treatment should begin with the standard starting dose, with the note that an increase in the dose will likely not be necessary. If side effects occur or the onset of action is delayed, the dose should be reduced in both poor and intermediate metabolizers, and ongoing efficacy should be monitored. In ultra-rapid metabolizers, attention should be paid to potential ineffectiveness. No interaction was found for methylphenidate with either CYP2D6 or COMT; the widespread assumption that COMT variants influence the response to MPH is therefore not confirmed here. No CYP2D6 interaction was found for clonidine either, making it a suitable alternative for patients with differing CYP2D6 status.¹²

This was confirmed in a clinical study involving 385 children aged 6 to 16 years, with 515 atomoxetine concentration measurements. After dose adjustment, intermediate metabolizers achieved concentrations 1.4 to 2.2 times higher than normal metabolizers and responded more frequently (93.6% versus 85.7%). With a single morning dose, a peak concentration of 268 ng/ml or higher was associated with a better response. Neurological side effects occurred more frequently at levels of 465 ng/ml or higher, and gastrointestinal side effects at 509 ng/ml or higher. This indicates that there is only a narrow window between the efficacy threshold and the side-effect threshold, which supports a stepwise titration approach for atomoxetine as well.¹³

1.1.2. Rapid Metabolizers of Atomoxetine: Up-dosing¶

For children with an activity value of 1.0 or higher—that is, ultra-rapid, normal, and intermediate metabolizers—the guideline uniformly starts with 0.5 mg/kg/day and increases the dose to 1.2 mg/kg/day after three days. If there is no effect after two weeks and no side effects are present, a peak concentration measurement is taken one to two hours after administration; if the level is below 200 ng/ml, the dose is increased proportionally toward 400 ng/ml. For an activity level of 0.5, treatment begins at 40 mg/day; if there is no response after two weeks, the dose is increased to 80 mg/day, and a measurement is taken two to four hours after administration. For adults, an intermediate step is added: 40 mg, then 80 mg after three days, then 100 mg if there is no response after two weeks, with blood level measurement taking place only after another two weeks. For ultra-rapid metabolizers, doses exceeding 100 mg/day may be necessary. The rationale for normal metabolizers is noteworthy: they respond less frequently than slow metabolizers and are more likely to discontinue treatment due to lack of efficacy. A common concern is overdose in slow metabolizers. The guideline identifies underdosing in fast metabolizers as the more common problem. All recommendations are classified as moderate strength. ¹¹

1.2. Dose-escalation rate for atomoxetine¶

Two controlled studies directly compare the rate of dose escalation for atomoxetine. In a nine-week, placebo-controlled study involving 181 children and adolescents aged 6 to 17, a rapid dose escalation regimen (0.5 mg/kg for 7 days, then 1.2 mg/kg) with a slow up-dosing regimen (0.5 and 0.8 mg/kg for 7 days each, then 1.2 mg/kg). There was no difference in efficacy, but the slower up-dosing was better tolerated.¹⁴

1.3. Target dose of atomoxetine¶

(E 4): For long-term use, the recommended daily dose is 1.2 mg/kg, and the maximum daily dose is 100 mg.(E 4)³ European treatment guidelines recommend starting with approximately 0.5 mg/kg/day for children and adolescents weighing up to 70 kg and increasing to the target dose of approximately 1.2 mg/kg/day no sooner than 7 days later. Neither 1.8 mg/kg/day nor 100 mg/day should be exceeded. (E 4)¹⁵

1.4. Time to response with atomoxetine¶

(E 4): The first effects may become apparent after a few days. It takes several weeks for atomoxetine to reach its full effect. Estimates vary: Dodson cites 8 to 10 weeks (E 4)¹, while European treatment guidelines suggest 6 to 8 weeks or longer, though those who respond typically show initial improvement starting in week 4. (E 4)¹⁵

In a pooled analysis of three open-label studies involving 338 children aged 6 to 11, the median time to at least a 25% improvement was 3.7 weeks. The median time to remission was 14.3 weeks. A significant improvement of at least 40% was observed in 47% of patients after 4 weeks at the latest, 76% after 12 weeks, 85% after 26 weeks, and 96% after 52 weeks. The likelihood of improvement therefore continues to increase over the course of several months. Anyone who considers atomoxetine ineffective after six or eight weeks is discontinuing treatment too early for a significant proportion of people with ADHD.¹⁶

A review of 125 studies on children and adolescents aged 6 to 18 reaches the same conclusion. Initial responses can occur quickly—within a week—or may take several weeks; in a six-week study, the median time was 23 days. The response builds up gradually and is often not reliable until after more than three months. It takes longer to reach full effect. The effect sizes compared to placebo ranged from 0.6 to 1.3 after 6 to 18 weeks. Six randomized studies involving 618 people with ADHD show that the response after four weeks predicts the response after six to nine weeks. Open-label data from 338 people with ADHD suggest that the probability of a significant response (at least a 40% reduction) continues to increase even beyond that point. The randomized comparisons with methylphenidate lasted only three to twelve weeks, which puts atomoxetine at a disadvantage due to its slow onset of action. ¹⁷ Several authors are employed by Eli Lilly, the manufacturer of atomoxetine.

1.5. Time of Day to Take Atomoxetine¶

(E 1b): ATX is usually taken in the morning. If fatigue occurs as a side effect, ATX can also be taken in the evening. (E 4)⁷ A single evening dose of ATX is associated with reduced efficacy but also with fewer side effects (E 1b)¹⁸

According to the prescribing information for Strattera, if people with ADHD experience side effects with a single daily dose of atomoxetine, the dose may be split into two equal doses taken in the morning and evening. (E 4)¹⁹

In a three-arm randomized trial involving 288 children aged 6 to 12 years, morning dosing was superior to evening dosing for some measures of efficacy, whereas evening dosing was significantly better tolerated. Significantly fewer children reported at least one side effect. Both dosing times were superior to placebo in terms of evening behavior. In cases of tolerability issues, switching to evening dosing is therefore an alternative to reducing the dose.²⁰In a randomized, double-blind trial, 288 children were assigned to morning atomoxetine (n = 102), evening administration (n = 93), or placebo (n = 93) in a randomized, double-blind trial lasting approximately six weeks, with a mean final dose of 1.25 and 1.26 mg/kg/day, respectively. Both dosing times reduced core symptoms compared to placebo, with effects still measurable 24 hours after administration. On some scales, morning dosing was superior, including the ADHD-RS total score, hyperactivity, impulsivity, and severity rating. There was a clear difference in tolerability. At least one adverse effect was reported by 74.0% of those receiving the morning dose, compared with 48.9% of those receiving the evening dose and 43.5% of those in the placebo group (p < 0.001 for each comparison with the morning dose). Evening dosing thus did not differ from placebo. Abdominal pain, decreased appetite, vomiting, drowsiness, nausea, and stomach discomfort were more common with morning dosing than with placebo. No single adverse effect differed significantly between morning and evening dosing; only the overall frequency did. Funded by Eli Lilly, the manufacturer of atomoxetine; several authors are employed by the company or were compensated as investigators.

The timing of the effect throughout the day was assessed in 421 children and adolescents aged 6 to 17 years over a 24-week period. With morning dosing (target dose 0.5 to 1.2 mg/kg/day), the evening score dropped from 13.7 to 8.0 as early as week 2, and the morning score decreased from 4.3 accordingly. The greatest change therefore occurs in the first few weeks, after which the effect continues to build more slowly. The most common side effects were fatigue (22.3%) and nausea (13.5%). Treatment was most frequently discontinued due to lack of efficacy (12.4%).²¹

1.6. Dose Monitoring for Atomoxetine¶

Since atomoxetine can increase blood pressure and heart rate, these should be monitored. (E 4)¹

1.7. Dosing of Atomoxetine¶

In our opinion, when titrating the dose—as with all antidepressant-like medications—a gradual approach is highly recommended. A different view is presented by Prasad et al. (E 4)⁴; however, we have received several reports of depression as a side effect of rapid discontinuation of atomoxetine.

2. Dosage of Guanfacine¶

Guanfacine appears to reduce norepinephrine levels during the first 2 weeks of treatment. Only after that does norepinephrine appear to increase due to receptor downregulation. This could have consequences for the full effect taking several weeks to set in—similar to antidepressants. (E 2b)²² Regardless, the typical side effects of guanfacine (fatigue, drop in blood pressure, slowed heart rate) occur particularly during the first few weeks of treatment. (E 1b)²³ (E 1b)²⁴ People with ADHD should therefore not assume that nothing is happening during this time.

 

According to the manufacturer, guanfacine is not dosed according to a fixed schedule but rather based on body weight. The target range is 0.05 to 0.12 mg/kg/day. Treatment begins with 1 mg daily, with the dose increased by no more than 1 mg per week. Doses exceeding 4 mg/day have not been studied in children aged 6 to 12 years, and doses exceeding 7 mg/day have not been studied in adolescents aged 13 to 17 years. The tablets must not be broken, chewed, or crushed, and they should not be taken with high-fat meals, as this increases the absorption of the active ingredient.²⁵

In a randomized, placebo-controlled study of adolescents aged 13 to 17 years (n = 314), guanfacine extended-release was titrated over a titration phase lasting several weeks to an individual dose between 0.05 and 0.12 mg/kg/day, up to a maximum of 7 mg/day depending on body weight. Discontinuations due to lack of efficacy were significantly less frequent in the active treatment group than in the placebo group (9 versus 25), whereas discontinuations due to side effects were more frequent (9 versus 3).²⁶

(ADxS experience): We are familiar with several patients for whom the dosage was adjusted individually based on symptom improvement and to avoid inappropriate side effects. This also seems more sensible to us.

Because of guanfacine’s blood pressure-lowering effect, it must be tapered off gradually when discontinuing treatment to avoid blood pressure problems. (E 4)²⁷

3. Adding additional non-stimulants¶

With other non-stimulant medications, such as bupropion, imipramine, or desipramine, it may also take 8 to 10 weeks for the full effect to take hold. (E 4)¹

3.1. Viloxazine Extended-Release (Qelbree)¶

Viloxazine extended-release (Qelbree) has been approved in the U.S. since 2021 for ADHD in children aged 6 and older, but not in Europe. The dosage is increased in fixed weekly increments. Children ages 6 to 11 start with 100 mg daily and increase the dose by 100 mg each week up to a maximum of 400 mg. Adolescents aged 12 to 17 start with 200 mg and, after one week, increase the dose by 200 mg in a single step to a maximum of 400 mg. Adults start with 200 mg and increase the dose by 200 mg weekly to a maximum of 600 mg. In cases of severe renal insufficiency, treatment should begin with 100 mg, with a maximum dose of 200 mg. Steady state is reached after just two days; no accumulation occurs. Heart rate and blood pressure should be monitored before starting treatment, after each dose increase, and throughout the course of treatment. Additionally, patients should be monitored for the emergence of suicidal thoughts.²⁸

3.2. Centanafadin (Simtriyo)¶

Centanafadin (Simtriyo) is a non-stimulant medication that has been approved in the U.S. since July 2026 for people with ADHD aged 6 and older who weigh at least 20 kg. It is a combined norepinephrine, dopamine, and serotonin reuptake inhibitor. The dosage for children is based on body weight and is otherwise largely standardized:²⁹³⁰

• Children: based on weight
  • 20 to less than 35 kg: 140 mg
  • 35 to 50 kg: 210 mg
  • over 50 kg; 280 mg
• Adolescents: 280 mg
• Adults
  • Start with 210 mg
  • Increase to 280 mg

Take in the morning, with or without food, at the same time each day if possible. The capsules must not be split, crushed, or chewed, but they can be opened and the contents added in their entirety to a tablespoon of applesauce or yogurt or mixed into orange juice; consume immediately and follow with water. The granules must not be chewed or stored.

Heart rate and blood pressure should be monitored before starting treatment, after each dose increase, and throughout the course of treatment. A warning box advises monitoring children for suicidal thoughts.

In the studies, the effects began within the first week—faster than with the other non-stimulants.

The pooled effect size across five studies is 0.37, which is in the range of atomoxetine and viloxazine. The drug is not approved in Europe. (As of August 2026)

In a randomized trial involving 459 adolescents aged 13 to 17, only the higher dose of 328.8 mg met the primary endpoint (mean change in the ADHD-RS-5 of -18.5 versus -14.2 with placebo; p = 0.0006), whereas the 164.4 mg dose did not differ significantly from placebo. The difference was already apparent as early as week 1.³¹

Before treatment begins, the approval text requires a check for

• Risk of Abuse and Dependence
• Heart disease, including personal and family history of sudden cardiac death and arrhythmias,
• Heart rate and blood pressure
• Risk factors for a manic episode
• Family history of tics or Tourette syndrome

Contraindications include

• Hypersensitivity
• concurrent use of a monoamine oxidase inhibitor or use within the past 14 days
• Pheochromocytoma
• structural heart defects
• Heart muscle disease
• serious heart rhythm disorders
• coronary heart disease

3.3. Sustained-release clonidine (Kapvay)¶

For sustained-release clonidine (Kapvay), dosing begins with 0.1 mg at bedtime for one week, followed by a weekly increase of 0.1 mg/day until the desired response is achieved, divided into two doses with an equal or higher evening dose. Doses above 0.4 mg/day have not been studied for ADHD and are not recommended. The tablets must not be split, chewed, or crushed, and they are not interchangeable milligram-for-milligram with other clonidine preparations.³²

When adjusting the dose, it should be reduced by no more than 0.1 mg every 3 to 7 days to avoid rebound hypertension. There are no studies on this for children with ADHD; the recommendation is based on the treatment of hypertension in adults.³²

Sustained-release clonidine suspension (Onyda XR). Approved in the U.S. since May 2024.
Onyda XR is the first liquid non-stimulant treatment for ADHD and the only non-stimulant that can be taken in the evening. Approved for use in children 6 years and older, either as monotherapy or in combination with stimulants.

Start with 0.1 mg (1 ml) at night, increasing the dose by 0.1 mg each week.
The prescribing information explicitly advises slow titration and frequent monitoring of vital signs, as dose-dependent decreases in blood pressure and heart rate may occur. In the underlying studies with sustained-release clonidine tablets (256 people with ADHD, two eight-week placebo-controlled studies), 75.5% were titrated up to the maximum dose of 0.4 mg/day.³³ Accordingly, the majority required the full dose
Onyda XR is approved for use as monotherapy or in combination with stimulants.

The suspension contains 0.1 mg per milliliter and should be shaken gently before use; this allows for more precise dosing than the tablet and is suitable for people with ADHD who are unable to swallow tablets.

Significant side effects:

• dose-dependent decreases in blood pressure and heart rate
• Sedation and drowsiness

Heart rate and blood pressure should be measured before starting treatment.

Its efficacy is not based on our own studies, but rather on data regarding sustained-release clonidine in tablet form; information on drug interactions is based on immediate release clonidine. In the event of an overdose, the effect occurs in two phases. The consequences of an initial rise in blood pressure include a drop in blood pressure, a slowed heart rate, respiratory depression, hypothermia, drowsiness, diminished reflexes, weakness, irritability, and constricted pupils. An initial rise in blood pressure is therefore not a cause for relief.

3.4. Bupropion¶

For bupropion in adults with ADHD, an 8-week placebo-controlled study involving 162 people with ADHD is available, in which the dose was titrated up to 450 mg/day of the once-daily XL formulation. Response was defined as at least a 30% reduction in symptoms. Bupropion lowers the seizure threshold in a dose-dependent manner, which is why the dose must be increased gradually. The sustained-release formulations, which have lower peak concentrations, carry a lower risk of seizures than the immediate-release formulations.³⁴

3.5. Medications containing THC¶

When titrating the dose of THC-containing medications, a starting dose of 1 to 2.5 mg of THC per day is recommended. The dose should be increased gradually, by 1 to 2.5 mg every 3 to 5 days. The average daily dose is 10 to 20 mg of THC. (E 4)³⁵

3.6. Modafinil¶

Modafinil is dosed based on body weight for ADHD and adjusted within a few days, not in weekly increments.

In a seven-week randomized, double-blind study involving 190 children and adolescents aged 6 to 17 years, the dose was titrated over the first 7 to 9 days to 340 mg/day for patients weighing less than 30 kg and to 425 mg/day for those weighing 30 kg or more. After abrupt discontinuation without tapering, neither withdrawal symptoms nor symptom rebound were observed.³⁶

With modafinil, it makes a difference whether the daily dose is divided into one or two doses. In a four-week placebo-controlled study involving 248 children aged 6 to 13, a single 300-mg dose taken in the morning was compared to the same total dose divided into two doses (100/200 mg or 200/100 mg). The single 300-mg dose provided the most consistent improvement in symptoms. The divided doses had inconsistent effects. Insomnia was the only side effect that occurred significantly more frequently than with placebo (14% versus 2% in one of the divided-dose groups).³⁷

With flexible dosing, the range of modafinil doses determined to be optimal was between 170 and 425 mg per day. In this nine-week, double-blind study of 200 children and adolescents aged 7 to 17, the effect size (0.60 to 0.78) was similar to that of atomoxetine and lower than that of methylphenidate.³⁸

4. Serotonin reuptake inhibitors¶

(ADxS experience): The use of serotonin reuptake inhibitors should always be discussed with a doctor on an individual basis. Serotonin reuptake inhibitors are suitable for the treatment of (comorbid) depression.

With regard to ADHD, treatment with minimal doses (e.g., 2 to 5 mg of escitalopram) might be useful, at most, for addressing impulsivity in ADHD-HI. In our experience, this appears to have an immediate effect and is apparently not mediated by receptor downregulation or upregulation.

Furthermore, serotonergic medications are not appropriate for treating ADHD itself, particularly in cases of ADHD-I.

5. Taper off blood pressure medication gradually¶

(No evidence): Maintenance medications must be introduced gradually and tapered off gradually.

With serotonin reuptake inhibitors, this can take more than half a year to avoid side effects.