Prescribing Atomoxetine

Why Atomoxetine Matters

Atomoxetine is the first drug the FDA ever approved specifically for adult ADHD, and it remains the prototype non-stimulant. Its selling point is simple and durable: it treats ADHD without being a controlled substance. No DEA paperwork, no monthly-visit requirement to generate a new prescription, no diversion, no abuse potential, no "my script got stolen" phone calls. You can send refills like you would for an SSRI. For a large slice of patients (those with substance use histories, those you can't see monthly, those who don't want a stimulant, those for whom stimulants failed or can't be used), that alone earns it a place in your toolkit.

But let's be honest about what it is and isn't. Atomoxetine is not as effective as a stimulant, and this isn't close. In head-to-head trials, stimulants post effect sizes of 0.8–0.9 and numbers-needed-to-treat of 2–4; atomoxetine lands at effect size 0.4–0.6 and NNT 5–7. In the largest pediatric head-to-head (Newcorn 2008, n=516), Concerta produced a 56% response rate versus atomoxetine's 45% versus placebo's 24%. It works, it clearly beats placebo, but it is a second- or third-line agent for garden-variety ADHD.

Part 1: Indications: Who Is Atomoxetine For?

FDA-Approved Uses

• ADHD in children age ≥6 and in adults: monotherapy, or as an adjunct to a stimulant.

That is the entire FDA label. Atomoxetine was originally developed as an antidepressant and failed; it found its second life in ADHD, where it became, in 2002, one of the first genuinely new ADHD mechanisms in decades.

Where Atomoxetine Earns Its Keep

This drug is best understood not as a first-line ADHD agent but as the answer to specific clinical situations:

ADHD + substance use disorder, the flagship indication

Because atomoxetine increases dopamine in the prefrontal cortex but not in the striatum, it has no reinforcing/euphoric effect and no abuse potential. It is unscheduled. For a patient with active or historical addiction, where handing over a bottle of amphetamine is fraught, atomoxetine is arguably the single best ADHD choice available. This is the case where atomoxetine may genuinely be your first choice, not a fallback.

ADHD + anxiety

Stimulants can wind anxious patients tighter. Atomoxetine tends not to, and there's real (if industry-sponsored) evidence it helps. A 14-week RCT in adults (Adler 2009, n=442) found atomoxetine beat placebo on both ADHD symptoms (p<0.001) and social anxiety (p<0.001). It's a reasonable pick for the ADHD patient whose comorbid anxiety you'd rather not aggravate.

When a controlled substance is undesirable or impractical

Patients who don't want a stimulant, adolescents in households where diversion is a concern, patients you can only see infrequently, athletes or professionals subject to drug testing, or anyone for whom the controlled-substance logistics are a barrier.

When stimulants failed or aren't tolerated

Insomnia, appetite suppression, tics, rebound irritability, cardiovascular jitteriness: atomoxetine sidesteps several of these and is worth a try when stimulants can't be optimized.

Off-Label Uses With Some Evidence

These are real signals, mostly from small trials, useful when the comorbidity coexists with ADHD, rarely worth chasing on their own:

• Binge eating disorder: 10-week RCT (n=40) reduced binge frequency (p=0.034) and weight (p=0.018).
• Obesity: 12-week RCT (n=20) reduced weight (p<0.0001).
• Obstructive sleep apnea: As a single agent (75–80 mg QHS) it cut the apnea-hypopnea index by roughly 50% versus placebo, and it's being developed in combination with agents like oxybutynin. Worth remembering because childhood ADHD carries a 20–30% rate of OSA, so the ADHD + OSA patient gets a two-for-one.

Where NOT to Use It

• Depression. Despite being born as an antidepressant, atomoxetine failed rigorous double-blind trials. If your patient has ADHD + depression, reach for bupropion or (for ADHD + depression) viloxazine instead, drugs that actually treat both.
• Bipolar disorder. This is a genuine caution, not a formality. Atomoxetine is noradrenergic, and noradrenergic agents (think venlafaxine, tricyclics) carry the highest risk of flipping patients into mania. Its exact switch rate is unknown, but there are case reports of mania and psychosis. In a patient with bipolar ADHD, the alpha-2 agonists (guanfacine, clonidine) are far safer choices.

Part 2: Before You Start: Workup and Candidacy

Atomoxetine requires far less pre-start machinery than lithium or a stimulant workup, but a few things are non-negotiable.

Baseline Assessment

• Baseline ECG is NOT routinely required. Atomoxetine can rarely prolong QTc, but the FDA does not recommend a routine baseline ECG. Get one if there's pre-existing cardiac disease, syncope, a concerning family history, or you're in an older/frailer patient.
• CYP2D6 genotyping is optional. The FDA leaves it to clinician judgment and does not recommend routine testing. It's most useful when you already suspect poor metabolism (severe side effects at low doses) or when you're stacking atomoxetine on a strong 2D6 inhibitor.

Who Is a Poor Candidate?

• Poorly controlled cardiovascular disease: the one true contraindication.
• Bipolar disorder (relative; high mania-induction concern).
• Narrow-angle glaucoma: noradrenergic agents can worsen it; use caution.
• Seizure disorder: caution (lowers threshold in theory).
• Tics: caution; monitor if pre-existing.
• Significant hepatic impairment: caution; dose-reduce.
• Patients who need a fast answer: if someone needs symptom control in days, not weeks, this is the wrong drug.

Part 3: How to Start and Dose

Formulation

Atomoxetine comes as capsules (available generic since 2017, so cost is no longer a barrier). A newer extended-release oral formulation can be opened and sprinkled, useful for children or anyone who can't swallow capsules. The capsules should not be opened onto the eye, since the powder is an ocular irritant, so counsel patients who sprinkle to avoid contact with the eyes and wash hands.

Despite a short (~5-hour) half-life, atomoxetine provides sustained daytime coverage with no rebound, and can be dosed once or twice daily. Once daily is simplest and usually sufficient; twice daily (morning and late afternoon) is a lever you can pull for GI intolerance or for patients who feel coverage fade.

Standard Dosing

Two ways to dose: by fixed milligrams (simple) or by weight (more precise, better for kids). Both aim at the same targets.

By fixed dose (typical adult/adolescent approach)

• Start: 40 mg/day.
• After ~3–7 days (minimum), increase to 80 mg/day.
• Assess after 2–4 weeks at 80 mg; if response is inadequate and it's well tolerated, go up to 100 mg/day (the usual maximum).

By weight (peds, and a good discipline in adults)

• Start: ~0.5 mg/kg/day (roughly 0.4–0.5 mg/kg).
• Titrate after a minimum of 3 days toward a target of ~1.2 mg/kg/day.
• Trials pushed to 1.8 mg/kg/day (or 100 mg, whichever is less); the average effective dose in comparative trials was about 53 mg/day.

The Two Rules That Actually Determine Success

1. Dose it high enough. See above. Target the milligram-per-kilogram number, not a comfortable low fixed dose.
2. Give it long enough. Set expectations on day one: "You may notice something in two weeks, but this drug keeps building for up to ten weeks. We won't judge it until you've had a solid four to six weeks at a full dose." This is the counseling that prevents premature discontinuation, the number-one reason atomoxetine "fails."

The CYP2D6 Dose Modification (Critical)

If your patient is a known CYP2D6 poor metabolizer (~2–10% of the population) or is taking a strong CYP2D6 inhibitor (fluoxetine, paroxetine, duloxetine, bupropion, or sertraline ≥150 mg/day):

• Start at the lowest possible dose.
• Titrate at half the usual speed.
• Target a dose 50–75% lower than usual.

In poor metabolizers, atomoxetine peak levels run roughly 5-fold higher (up to ~16-fold in some reports), and these patients are twice as likely to discontinue for side effects. This is the closest thing atomoxetine has to a narrow-therapeutic-window problem, and it's entirely manageable if you check the med list before you write.

Part 4: Monitoring

Atomoxetine has no serum level to chase and no routine bloodwork mandate. Monitoring is clinical and cardiovascular.

• No routine labs, no routine ECG, no drug levels.
• Recheck an ECG only if new palpitations, syncope, or you've stacked another QTc-affecting drug.
• In children and adolescents, monitor for suicidal ideation the way you would when starting an antidepressant, more closely in the first few weeks and after dose changes.
• Watch for mania/psychosis in anyone with a bipolar diathesis.

Part 5: Side Effects and How to Manage Them

Most atomoxetine side effects are dose-related, worst during titration, and improve with time. The governing move is to titrate slowly enough that the patient stays on the drug long enough for it to work.

Gastrointestinal (Nausea, Upset Stomach)

Common, especially early.

Management:

• Take with food: the simplest and most effective fix.
• Split to twice daily to lower peak concentration.
• Slow the titration; most GI upset settles within a couple of weeks.

Fatigue and Somnolence: the Signature Nuisance

This is atomoxetine's most distinctive tolerability problem and the mirror image of stimulants (which cause insomnia). Roughly 1 in 20 patients (5%) get severe fatigue; atomoxetine ranks 2nd among 30 antidepressant-class drugs for FDA-reported severe fatigue. It can be genuinely disabling.

Management:

• Move the dose to the evening if daytime somnolence dominates, since the short half-life means much of the sedation can be slept off.
• Conversely, if it causes insomnia (it can go either way), dose in the morning. Let the side-effect direction pick the timing.
• Take with food.
• If fatigue is severe and persistent, it may simply be the wrong drug for this patient, so don't force it.

Appetite and Weight

Appetite suppression occurs but is less than with stimulants. Some modest weight loss in kids; usually self-limited. Track growth in children as you would on any ADHD drug.

Cardiovascular

Atomoxetine raises heart rate and blood pressure modestly, just like a stimulant, via its noradrenergic action.

Management:

• Recheck HR/BP at each dose step and periodically.
• If BP or HR climbs meaningfully, reduce the dose or reconsider the drug; coordinate with the PCP if the patient is on antihypertensives.
• Rare QTc prolongation exists but does not warrant routine ECG screening.

Psychiatric / CNS

• Suicidal ideation: carries an antidepressant-style warning in children and adolescents (see the boxed warning in Part 6). Case reports exist; it was not statistically significant in controlled trials. Monitor early, as with any antidepressant start in youth.
• Mania / psychosis: rare, via the noradrenergic mechanism. Real risk in bipolar-spectrum patients, so screen first, watch after.
• Irritability, mood lability: occasional; usually dose-related.

Genitourinary / Sexual (Adults)

Noradrenergic tone can cause urinary hesitancy/retention and sexual dysfunction (and, uncommonly, priapism; counsel patients to seek urgent care for a prolonged erection). These are worth asking about directly, since patients rarely volunteer them.

Hepatic (Rare but Important)

Part 6: Overdose, Toxicity, and Warnings

Atomoxetine has no boxed warning for the classic toxicity syndromes and no narrow therapeutic window in the lithium sense. Its safety profile is closer to an antidepressant's than a stimulant's.

The Boxed Warning

Other Warnings

• Rare sudden cardiac death: a signal shared with stimulants (FDA warning era, 2006), driven by underlying structural/electrical heart disease. This is why poorly controlled cardiovascular disease is a contraindication and why you screen the heart before starting.
• Hepatotoxicity: rare, idiosyncratic; stop for any sign of liver injury.

Overdose

Overdose experience is limited and generally less lethal than with stimulants or tricyclics. Expected features reflect exaggerated noradrenergic effect: tachycardia, hypertension, agitation, somnolence, GI upset, and occasionally seizures or QTc changes. Management is supportive; there is no specific antidote, and because atomoxetine is highly protein-bound, dialysis is not useful. Any significant overdose is an ED evaluation with cardiac and neurologic observation.

Part 7: Drug Interactions: It's All About CYP2D6

Atomoxetine is metabolized through essentially one hepatic pathway: CYP2D6. A single-pathway drug is a drug whose levels swing hard when you touch that pathway. This is the interaction chapter that matters; everything else is minor.

Strong CYP2D6 Inhibitors (Raise Atomoxetine Levels)

Co-prescribing any of these turns your patient, pharmacologically, into a poor metabolizer:

• Fluoxetine
• Paroxetine
• Duloxetine
• Bupropion
• High-dose sertraline (≥150 mg/day)

Action: start at the lowest dose, titrate at half speed, and target 50–75% below the usual dose. These are common psychiatric co-meds; an ADHD patient on fluoxetine or bupropion is an everyday scenario, so this is not an edge case. Check the list every time.

CYP2D6 Poor Metabolizers (Genetic)

Same clinical picture without any inhibitor: ~5-fold higher peaks, twice the discontinuation rate. Same dosing strategy (lowest start, half-speed, 50–75% lower target). Genotyping is optional but can explain a patient who can't tolerate even small doses.

Combinations to Avoid or Approach Cautiously

• Atomoxetine + stimulant: safety "not fully established." Some reports support it, but if you need to augment, guanfacine has actual controlled-trial evidence for augmenting stimulants with fewer cardiovascular concerns. Prefer the alpha-2 agonist.
• Other noradrenergic/pressor agents (including MAOIs; avoid the combination, as with any noradrenergic drug) can compound hypertension and mania risk.

Part 8: Special Populations

Pregnancy

Human safety data are limited; risk is not well established. There is no signal as alarming as valproate's, but there's also not enough data to call it safe. Weigh the severity of ADHD against the uncertainty (many patients can pause ADHD pharmacotherapy through pregnancy) and make an individualized, documented decision, ideally with maternal-fetal input.

Lactation

Safety unknown; atomoxetine is expected to enter breast milk. If used, monitor the infant for sedation, poor feeding, and irritability, and involve pediatrics.

Children and Adolescents (≥6 years)

• Approved from age 6. Dose by weight (0.5 mg/kg start → ~1.2 mg/kg target, max 1.8 mg/kg or 100 mg).
• Effect size 0.4–0.6.
• Suicidal ideation warning applies: monitor early.
• Remember the 20–30% OSA comorbidity: the ADHD + sleep-disordered-breathing child may get a bonus.
• Track growth and vitals.

Adults

Approved and effective, effect size ~0.45. Same CYP2D6 cautions. The adult ADHD patient with anxiety, a substance-use history, or an aversion to controlled substances is the classic adult atomoxetine candidate.

Elderly

Little specific data. Use extra caution for the cardiovascular effects (HR/BP), consider a baseline ECG in the frail, start low, and reconcile the (often long) medication list for 2D6 inhibitors.

Hepatic Impairment

Because clearance is hepatic, reduce the dose in hepatic impairment (moderate impairment generally warrants ~50% reduction; severe, ~75%). No routine adjustment is needed for renal impairment.

Renal Impairment

No specific dose adjustment is generally required; atomoxetine is not renally cleared to a clinically limiting degree.

Part 9: Discontinuation

This is one of atomoxetine's genuine conveniences. Atomoxetine has no discontinuation syndrome and no rebound, unlike stimulants (which can rebound within hours) and unlike the alpha-2 agonists clonidine and guanfacine (which risk rebound hypertension on abrupt stop and must be tapered).

• You can stop atomoxetine without a taper from a physiologic standpoint.
• A brief taper is still reasonable clinical courtesy (it lets you watch for return of ADHD symptoms and any mood change), but it isn't pharmacologically required.
• There is no "lose the response by stopping wrong" phenomenon here, as there is with lithium.

Part 10: Atomoxetine vs the Alternatives

vs Stimulants (Methylphenidate, Amphetamines)

Stimulants win on efficacy, and it isn't close: effect size 0.8–0.9 vs 0.4–0.6; NNT 2–4 vs 5–7; 56% vs 45% response in the key head-to-head. Stimulants also work in hours, not weeks. Atomoxetine's counter-advantages: not controlled, no abuse potential, once-daily coverage with no rebound, safer in substance use, and better for comorbid anxiety. Stimulants cause insomnia and appetite loss; atomoxetine causes fatigue. For most straightforward ADHD, a stimulant is the better first drug; reserve atomoxetine for the specific niches.

vs Viloxazine (Qelbree)

The other unscheduled non-stimulant NRI, and a close analog. Similar efficacy (NNT ~5–7). Viloxazine adds serotonergic activity, lacks the sudden-cardiac-death warning, and reaches full effect a bit faster (4–6 weeks vs up to 10). Atomoxetine is more studied and longer on the market. Reasonable to try one if the other disappoints.

vs Alpha-2 Agonists (Guanfacine ER/Intuniv, Clonidine ER/Kapvay)

Different mechanism entirely (alpha-2A agonism, not reuptake inhibition). Comparable-to-slightly-different effect sizes (guanfacine 0.4–0.9, clonidine 0.7–0.8). Atomoxetine is generally better tolerated and helps anxiety; the alpha-2 agonists improve sleep, reduce tics, and are safer in bipolar disorder, and guanfacine has the better evidence for augmenting a stimulant. Pick by comorbidity: tics/insomnia/bipolar → alpha-2 agonist; anxiety/substance use → atomoxetine.

vs Bupropion (Off-Label)

Bupropion is weaker for ADHD (effect size 0.3–0.5) but is the go-to when depression or nicotine dependence rides along. Note it's also a 2D6 inhibitor if you ever combine.

Mechanism: The Short Version

Atomoxetine is a selective norepinephrine reuptake inhibitor (NRI). By blocking the presynaptic norepinephrine transporter, it raises synaptic norepinephrine, and, because the prefrontal cortex clears dopamine largely via the NE transporter, it also raises dopamine in the frontal lobes. Crucially, it does not raise dopamine in the striatum/nucleus accumbens, the reward circuitry that mediates euphoria and abuse. That single anatomical fact explains atomoxetine's entire personality: it treats the prefrontal attention/executive deficits of ADHD, but it isn't reinforcing, isn't abusable, and isn't scheduled.

Two consequences follow directly from the mechanism. First, the noradrenergic load is why atomoxetine raises HR/BP and carries a mania-induction concern; it's pharmacologically a noradrenergic antidepressant wearing an ADHD label. Second, its ~5-hour half-life through a single CYP2D6 pathway is why the 2D6 story dominates its interactions and why poor metabolizers get hit so hard. It shares its core mechanism with reboxetine, the European antidepressant, and viloxazine, its newer non-stimulant sibling.

The Bedside Cheat Sheet