Non-stimulants for adult ADHD

The five options at a glance

Two non-stimulants are US Food and Drug Administration (FDA) approved in adults: atomoxetine(Strattera, since 2002; generic since 2017) and viloxazine ER(Qelbree, since April 2022). Two are pediatric-approved and adult off-label but widely used: guanfacine ER (Intuniv) and clonidine ER (Kapvay). One is fully off-label across all indications for ADHD: bupropion(Wellbutrin). That’s it. Don’t let the “five FDA-approved non-stimulants” framing in pop content mislead you — it’s two on-label adult, two adult off-label, one fully off-label.

The foundational efficacy citation underneath everything is Cortese et al. (2018) Lancet Psychiatry network meta-analysis¹ of 133 double-blind trials. Standardised mean differences on clinician-rated outcomes in adults: amphetamines ~0.79, methylphenidate ~0.49, atomoxetine ~0.45. Modafinil is around 0.16 and not approved for ADHD. The non-stimulants are not the consolation prize; they sit roughly where methylphenidate sits. The meaningful difference between non-stimulants and stimulants is less about effect size than about onset speed, side-effect profile, and what they treat alongside the core ADHD symptoms.

Atomoxetine — the patience tax

Selective norepinephrine reuptake inhibitor. Adult dose 40–100 mg/day, usually titrated to 80 mg over 2–4 weeks. The one fact most prescribers get wrong when they hand it to you: onset is slow. First effects 2–4 weeks. Full effect 6–12 weeks. Patients who quit at week 2 are quitting before the drug has had a chance. Cunill et al. (2016) meta-analysis² confirms modest but reliable efficacy vs placebo, with smaller effects than stimulants and slower onset. The label flags a black-box warning for suicidal ideation in children and adolescents; the adult evidence didn’t replicate the signal but the warning is on the bottle.

Side effects: nausea (worst in the first two weeks, usually tolerable if you take it with food), sexual dysfunction (real, often under-warned about), fatigue, modest blood pressure and heart rate elevation requiring routine monitoring. Sweet-spot patient: stimulant-intolerant; anxiety comorbid (atomoxetine doesn’t worsen anxiety the way stimulants sometimes do); cardiac contraindication; substance-misuse history; or a job that screens for controlled substances. Where it doesn’t fit: someone who needs a same-day functional shift, or someone who can’t hold a stable medication regime for two months while waiting for it to work.

UK-specific: the National Institute for Health and Care Excellence (NICE) guideline NG87 lists atomoxetine as a second-line option after stimulants. Some integrated services offer it as a starter for patients with addiction history or where stimulant access is constrained.

Viloxazine ER (Qelbree) — the new one

FDA-approved for adults in April 2022 — the first new ADHD molecule in over a decade. Multimodal serotonergic-noradrenergic agent: norepinephrine reuptake inhibitor plus 5-HT2B antagonism and 5-HT2C agonism. Not a classical serotonin-norepinephrine reuptake inhibitor (SNRI). Adult dose 200–600 mg/day. Onset around 2 weeks for first effects.

Nasser et al. 2022 adult RCT³ showed Adult ADHD Investigator Symptom Rating Scale (AISRS) reductions vs placebo — modest but statistically significant; placebo-adjusted reduction around 3.7 points on AISRS at the 600 mg dose. Most common side effects: insomnia, somnolence (paradoxical — some patients get one, some get the other), headache, nausea, fatigue, decreased appetite. CYP1A2 substrate and inhibitor— meaningful interactions with caffeine, theophylline, clozapine, duloxetine, melatonin, and tizanidine. Patients on heavy daily caffeine need to know this; viloxazine can substantially raise caffeine plasma levels, producing what looks like an “anxiety reaction” that’s really an interaction. Adult label also carries a pediatric-trial black-box warning for suicidal thoughts.

Sweet-spot patient: someone who hasn’t tolerated atomoxetine (different mechanism, different side-effect profile); someone with comorbid anxiety where the 5-HT2 modulation may help; someone who wants a non-stimulant that’s not 20 years old. Where it doesn’t fit: anyone on the CYP1A2 interaction list above who can’t adjust caffeine or co-medications.

Guanfacine ER and clonidine — the α2A agonists

Both are α2-adrenergic agonists working on prefrontal-cortex postsynaptic receptors — Arnsten’s mechanistic work⁶ in non-human primates established the executive-function effect. Guanfacine is the more α2A-selective of the two; clonidine hits more α2 subtypes and is older, more sedating, more hypotensive. Both are FDA-approved in pediatric ADHD; both are widely used off-label in adults.

Guanfacine ER (Intuniv): adult dose typically 1–4 mg/day, titrated by 1 mg/week. Iwanami et al. (2020) Japanese adult phase 3 trial⁴ showed significant Conners’ Adult ADHD Rating Scale (CAARS) reduction at 4–6 mg/day — the strongest single adult RCT for guanfacine. Side-effect profile: sedation (especially in the first two weeks), orthostatic hypotension, bradycardia, dry mouth. Do not stop abruptly — rebound hypertension is a real risk. Best fit: hyperarousal, emotional dysregulation, sleep-onset disruption, comorbid tics, or as adjunct to a stimulant for evening coverage.

Clonidine ER (Kapvay): adult dose typically 0.1–0.4 mg/day. Sparse adult RCT data — pediatric data extrapolated to adult prescribing. Primary adult use is for sleep-onset and hyperarousal rather than daytime executive function. More sedating than guanfacine; same warning on abrupt discontinuation.

Bupropion — the off-label adjunct

Off-label for ADHD at any age — FDA-approved for depression and smoking cessation. A norepinephrine-dopamine reuptake inhibitor (NDRI). Verbeeck et al. 2017 Cochrane review⁵ pooled six adult ADHD trials and found a modest effect, with low confidence and a wide range of results. Useful in three specific situations: comorbid depression where you want one medication doing two jobs; a stimulant-misuse history that contraindicates controlled substances; smoking cessation as a co-goal. Adult dose typically 150–450 mg/day XL formulation.

Two contraindications worth flagging plainly: active eating disorders and seizure history (bupropion lowers seizure threshold). Abrupt alcohol cessation also raises seizure risk on bupropion.

When a non-stimulant is genuinely first-line

Six situations move a non-stimulant from second-line to first. None of them are preference for “something milder.” They are specific clinical and structural reasons stimulants don’t fit.

• Cardiac contraindication — uncontrolled hypertension, recent myocardial infarction (MI), certain arrhythmias. Atomoxetine and guanfacine sit differently on the cardiovascular profile.
• Substance-misuse history— particularly stimulant misuse. Non-stimulants don’t carry abuse potential and don’t require Schedule II handling.
• Pregnancy and breastfeeding — stimulant safety data are mixed; some prescribers prefer non-stimulants by default in these windows. Discuss specifically with a prescriber.
• Military, security clearance, or commercial-aviation roles where Schedule II stimulant prescribing creates job complications.
• Severe rebound or peak/trough oscillationon stimulants that doesn’t resolve with formulation changes — an α2A agonist or atomoxetine offers smoother coverage.
• Anxiety-dominant comorbiditywhere stimulants worsen anxiety. Atomoxetine and viloxazine often don’t.

Combining stimulants + non-stimulants

Common in clinical practice; under-published in the adult literature. The available evidence base is mostly Joshi, Wilens, and Spencer work in pediatric and adolescent populations, extrapolated to adult prescribing. Two common combinations: stimulant + α2A agonist(guanfacine or clonidine for evening coverage when the stimulant has worn off — addresses sleep-onset, irritability, and emotional dysregulation that stimulants don’t reach); and stimulant + atomoxetine (where a stimulant produces peak/trough oscillation the patient finds intolerable and steady-state atomoxetine smooths the curve). Adult RCTs of these combinations are thin. Both are reasonable clinical practice; neither is heavily evidenced.

The “don’t work” myth and the prior-auth reality

The dominant community framing is that non-stimulants don’t work. The Cortese 2018 SMD hierarchy gives the seed of truth — amphetamines outperform — but a 0.45 SMD for atomoxetine is not “doesn’t work.” The pattern that produces the myth is usually one of three: people quit before the onset window, people get poorly titrated (atomoxetine started at full dose with no titration produces nausea most can’t tolerate; guanfacine started at 2 mg produces sedation severe enough to stop), or people are comparing to their own peak amphetamine response rather than to no treatment.

Insurance reality in the US: many plans now require a documented failed stimulant trial before authorising viloxazine, and some still require step-therapy on atomoxetine before guanfacine. UK NICE NG87 places non-stimulants as second-line, which structures NHS access accordingly. NICE NG87¹² and the AAFP 2020 adult-ADHD pharmacology review¹¹ both articulate the prescribing sequence cleanly. Worth reading either before the appointment.

What’s contested

Pediatric-to-adult extrapolation is the contested foundation under guanfacine and clonidine adult prescribing. Both were trialled and approved in children. Adult α2A receptor distribution differs from pediatric, and the adult ADHD RCT evidence is thinner than the labels imply — Iwanami 2020 is the strongest single adult guanfacine RCT and is a Japanese sample with regional regulatory implications. Adult use is mechanistically reasonable, off-label, and not as well-evidenced as the pediatric labels suggest.

Adult combination-therapy RCTs (stimulant + non-stimulant) are genuinely thin. Bupropion’s adult-ADHD evidence base is small and the Cochrane review flagged low confidence. Long-term comparative effectiveness studies (does atomoxetine over 5 years beat methylphenidate over 5 years on functional outcomes?) don’t exist at scale. The reasonable position is that the non-stimulants are an honest second-line, sometimes first-line option with a slower onset and a different side-effect tail — not the consolation prize, and not the equal substitute either.