Strattera, Qelbree, Intuniv, Kapvay, Wellbutrin, Modafinil. What each one does, who it fits, and why most patients should not switch off a working stim.
Non-stimulants exist for ADHD because stimulants don’t fit every patient, and any honest conversation about ADHD medication has to start there. The stimulant class is the most reliable tool we have for the core attention symptoms, and on average it wins the head-to-head trials by a meaningful margin, but “on average” hides the patient sitting in front of you who can’t take a stim, shouldn’t take a stim, or already tried one and got worse. The non-stim lane is built for those patients, and it earns its keep when the picture rules out the stim lane for cardiac reasons, for an anxiety pattern stims make uglier, for a history of substance misuse that makes a Schedule II prescription a bad idea on principle, or for the patient who had a bad response to the first three stims he tried. The class is also useful as add-on therapy with a stim, smoothing out the late-day “the Vyvanse wore off at 4 p.m. and life still requires me to think” hours.
The honest framing, before we get into the individual drugs, is that the non-stim class works less reliably for the core attention symptoms than stims do. Effect sizes in the trials are smaller, response rates are lower, and the time-to-effect is measured in weeks rather than hours. That isn’t a marketing problem with the class, that’s the actual pharmacology, and a prescriber who pretends otherwise is selling a story that won’t survive the first three months. When the non-stim fits, it can be the difference between a workable life and a stuck one. When it doesn’t fit, the patient takes a pill every day, waits eight weeks, and finds out he was waiting for nothing. Picking the right tool for the shape of the patient is the whole job here.
Strattera (atomoxetine), the original of the lane
Strattera (atomoxetine) was the first FDA-approved non-stim for ADHD, landed in 2002, and is still the workhorse of the lane. The mechanism is a selective norepinephrine reuptake inhibitor, putting a thumb on the molecular vacuum that pulls norepinephrine back into the sending neuron, leaving more of the neurotransmitter in the synaptic gap, and over weeks the brain adapts to the new steady-state level in the prefrontal cortex where the executive function work happens. The relevant difference from stims is that there is no dopamine hit at the reward circuits, which is why Strattera doesn’t have abuse liability and is not a controlled substance. The cost of that design is that the patient doesn’t feel a dose the way he feels a stim. There is no “the Adderall kicked in” moment, no daily on-and-off curve, no felt difference between day three and day four. The drug works by raising the baseline over weeks, and the patient notices the lift, if he notices it at all, between week four and week six.
Dosing usually starts at 40mg for one to two weeks and titrates up to 80mg, where most adult patients land, with 100mg as the ceiling for the harder cases. Going under 40mg in an adult is generally a waste of the patient’s time, the dose-response curve doesn’t get out of bed at the lower numbers, and a patient who has been on 25mg for two months and reports no benefit hasn’t actually had a real trial. The side effect picture is its own conversation. Sedation is common in the first couple of weeks and usually settles, dry mouth is common and doesn’t always settle, GI upset shows up in the first week and tapers, sexual dysfunction (delayed orgasm, reduced libido, erection trouble) is real and underdiscussed in clinic. Blood pressure and heart rate creep up a few points in most patients, smaller than the stim signal but real, and worth tracking with a home cuff in anybody with a cardiovascular baseline that matters. The boxed warning for suicidal ideation in adolescents and young adults sits on the label and a thoughtful prescriber takes it seriously without treating it as a deal-breaker in an adult with a stable history.
The single most common failure pattern with Strattera is the patient who quits at week two. The drug hasn’t done anything yet because the drug doesn’t do anything yet at week two, the side effects are at their peak, and the patient calculates that he is paying a side-effect cost for zero benefit. So he stops. The fix is to set that expectation on day one, in plain words, “you are not going to feel this drug the way you would feel a stim, the side effects are going to land before the benefit does, and the deal we are making is that you give it six weeks before you decide it doesn’t work.” Real talk, when Strattera works, it is genuinely good, the focus lift is steady and durable and doesn’t crash. But the work has to be done over weeks, not days, and the prescriber who pretends otherwise is setting the patient up to quit.
Qelbree (viloxazine ER), the newer cleaner cousin
Qelbree (viloxazine extended-release) got FDA approval in 2021 for kids and in 2022 for adults, and it sits in roughly the same chemical neighborhood as Strattera, a norepinephrine reuptake inhibitor with some additional serotonin receptor modulation thrown in. The interesting part of the molecule for the practicing prescriber is that the side effect burden is often cleaner than Strattera’s at comparable doses, and the time-to-effect is two or three weeks faster in clinical practice, which makes the wait less painful and the quit-at-week-two pattern less common. For the right picture, Qelbree is often a better lane than Strattera.
Dosing in adults starts at 200mg, climbs to 400mg, and tops out at 600mg, with most patients landing between 400mg and 600mg. The titration is faster than Strattera’s, the steady-state lift comes in between week two and week three for responders, and the side effect profile leans toward mild sedation and headache early on, with the sexual side effects and the GI burden landing softer than Strattera’s in most patients. Blood pressure goes up a few points, same lane as Strattera, worth tracking the same way.
The catch on Qelbree is cost. The drug is branded only, no generic yet, and the cash price without insurance is brutal. Most plans require failing Strattera and at least one stimulant before they cover Qelbree, and the prior authorization paperwork is its own afternoon. When the patient has the right picture and the right insurance posture, Qelbree gets approved and often beats Strattera on tolerability. For a lot of patients who would do well on Qelbree, the cost and the prior auth maze means they end up on Strattera instead, and that’s a system problem rather than a clinical one. The prescriber’s job is to make sure the patient at least knows the option exists.
Guanfacine ER (Intuniv), the wound-tight patient’s drug
Intuniv (guanfacine extended-release) started its life as a blood pressure medication, an alpha-2A adrenergic receptor agonist that dampens sympathetic outflow and drops systolic numbers a few points. The ADHD use, FDA-approved in 2009 for kids and off-label in adults for as long as anybody has been paying attention, comes at lower doses than the BP indication and targets the executive function piece without any dopamine work. The receptor lives in the prefrontal cortex where attention, working memory, and impulse filtering all happen, and stimulating it from the outside supports the same circuit work the brain’s own norepinephrine is supposed to be doing.
The fit is specific. The patient who responds best to Intuniv is the one with the wound-tight picture, high baseline arousal, low filter, irritability layered into the attention problem, not just inattentive but reactive, the one who blows up at small frustrations, can’t downshift after work, lies in bed at 11 p.m. with his nervous system still running at the speed of his afternoon. The combination of inattention and high arousal is the lane where Intuniv earns its keep, because it takes the edge off the arousal piece without sedating the patient into uselessness, and the focus lift comes from the same dose. Add it to a stim and it does double duty, smoothing out the stim hyperactivation (the buzzy edge, the teeth-grinding, the heart pounding at 2 p.m.) while extending the executive function support into the early evening.
Dosing starts at 1mg, climbs by 1mg per week as tolerated, and tops out at 4mg in adults. Going faster than that is asking for trouble. The dose-limiting side effects are sedation (early, before tolerance builds over two or three weeks), the blood pressure drop (real, dose-dependent, and a problem if the patient is already running low or on another antihypertensive), and the orthostatic dizziness at the higher doses. Dosing at bedtime helps with the sedation piece, which often turns into a sleep benefit, the wound-tight patient finally winding down at 10 p.m. instead of 12:30. Coming off Intuniv has to be a taper, never a sudden stop, because the abrupt withdrawal of an alpha-2 agonist can cause a rebound spike in blood pressure, the heart rate climbs, the patient feels jittery and headachy for a few days while the system resets. A taper over a couple of weeks at the same 1mg-per-step interval handles it cleanly.
The non-stim class is built for the patients stims don’t fit, and the fit conversation matters more in this lane than in any other class in psychiatry.
Clonidine ER (Kapvay), the older workhorse
Kapvay (clonidine extended-release) sits in the same alpha-2 agonist family as Intuniv, but it’s the older molecule, hits the receptors a little less selectively, sedates harder, and drops blood pressure more aggressively. FDA approval for ADHD came in 2010, the off-label use predates that by decades, and it is still a real tool in the right patient, especially the adult with the wound-tight picture plus sleep-onset trouble that the more sedating molecule actually helps with rather than ruins.
Dosing starts at 0.1mg, climbs to 0.4mg at the top end, and usually splits between morning and bedtime because the half-life is short enough that once-a-day doesn’t cover the day cleanly. Cost is where Kapvay has a real edge over Intuniv, generic clonidine is cheap, sometimes pennies-per-pill cheap, and for the patient whose insurance is fighting him on everything, it can be the move that gets a real medication into the picture without the financial fight. The combo with a stim is the same logic as Intuniv plus a stim. Less commonly first-line in adults now because Intuniv handles the same job with less sedation and less BP drop, but when the picture fits and the cost matters, Kapvay is a real workhorse.
Wellbutrin (bupropion) for ADHD, off-label but useful
Wellbutrin (bupropion) is not FDA-approved for ADHD, but it has been used off-label in the indication for years and it has a real evidence base in the adult ADHD literature that doesn’t get the airtime it deserves. The mechanism is a norepinephrine and dopamine reuptake inhibitor, the only antidepressant on the market that hits the dopamine system in a meaningful way, and the dopamine piece is what gives it the ADHD signal that the SSRIs and SNRIs don’t have. It isn’t going to match a stim for the core attention symptoms in most patients, but it can take a real bite out of the picture in patients where the rest of the conversation lines up to make it the smart pick.
The patients where Wellbutrin earns its place in the ADHD lane are usually carrying more than just ADHD. The picture is ADHD plus low-grade depression that doesn’t quite meet the bar for a separate diagnosis but is grinding on the patient anyway, or ADHD plus a nicotine vape habit the patient wants to quit (Wellbutrin is FDA-approved as Zyban for smoking cessation, same molecule covers both jobs), or ADHD plus the sexual side effect picture from a previous SSRI trial. Wellbutrin has almost no sexual side effects of its own, which patients appreciate. Dosing usually starts at 150mg of the XL form and climbs to 300mg, with 450mg available for harder cases.
The caveat that has to be on the table before the first script is written is the seizure threshold. Wellbutrin lowers it, the risk is dose-dependent, and there are specific patient groups where the math doesn’t work out, anybody with a history of seizure disorder, anybody with active eating disorder pathology (bulimia and anorexia both elevate seizure risk and the drug is contraindicated), anybody with active or recent alcohol withdrawal risk. In a patient without those risk factors, the absolute seizure risk at standard doses is small, under one percent, and the drug is a reasonable pick for the right picture. In a patient with those risk factors, the other non-stim options are usually the better lane.
Modafinil and armodafinil, the wakefulness lane
Modafinil (Provigil) and armodafinil (Nuvigil) are wakefulness agents, FDA-approved for narcolepsy, shift work sleep disorder, and obstructive sleep apnea as adjunct treatment, with a real but off-label use in ADHD when the stim lane is closed for other reasons. The mechanism is messy, the drugs hit the dopamine transporter at lower affinity than amphetamines do but in the same general neighborhood, with additional effects on orexin, histamine, and glutamate that the textbooks describe in slightly different fonts. The practical experience is that modafinil produces a focus and wakefulness lift somewhere between caffeine and a low-dose stim, without the same euphoria, without the same crash, and without the same abuse pattern.
The fit is the patient whose stim trials worsened anxiety in a way he couldn’t tolerate but whose cognitive picture still wants a chemical assist, the patient with a substance misuse history where the Schedule II prescription is a non-starter but the Schedule IV is on the table, the patient with shift work layered on top of his ADHD. Dosing is 100mg to 400mg for modafinil and 150mg to 250mg for armodafinil, taken in the morning, with both lasting most of the day on a single dose. Side effects are typically headache (often settles), insomnia if dosed too late in the day, low-grade anxiety in some patients, and the rare but serious Stevens-Johnson syndrome risk that lives on the label and that the patient has to know about before the first dose.
The insurance reality on modafinil and armodafinil is heavy. Off-label for ADHD means the prior authorization paperwork starts at no, and getting it to yes usually requires documenting failed stim trials, a specific reason the stim lane is closed, and sometimes a letter of medical necessity. Generic modafinil has been around since 2012 and is cheap on the discount card programs even when insurance won’t cover it, which is often the move when the prior auth maze isn’t worth the patient’s time. Armodafinil is more expensive and harder to get covered, and the head-to-head data between the two molecules isn’t strong enough to justify the cost gap for most patients.
The combo strategy, non-stim plus stim
Adding a non-stim to a stim is one of the most common moves in adult ADHD practice past age thirty, and the receptor math holds together once you look at it. The stim is doing dopamine and norepinephrine work at the reward and the executive function circuits, the alpha-2 agonist is doing receptor-level support on the executive function piece without touching dopamine, Strattera or Qelbree is doing pure norepinephrine reuptake inhibition on a different timescale, and Wellbutrin is hitting dopamine and norepinephrine through a reuptake mechanism instead of a release mechanism the way amphetamines do. Different angles into the same circuit. Stacking two of them often gives a smoother day than either alone, with the stim handling the core attention symptoms and the non-stim handling the edges (the after-stim hours, the wound-tight component, the late-evening executive function demands when the kids need dinner and the bills aren’t going to pay themselves).
The common pairings in practice are guanfacine plus Vyvanse for the patient who needs the wound-tight edge taken off, Strattera plus a methylphenidate product for the patient who wants more sustained focus support across a long day, and Wellbutrin plus Adderall for the patient carrying ADHD plus low-grade depression that the SSRI lane wouldn’t have addressed without wrecking his sex life. The dose math has to be worked out individually, the BP and HR have to be tracked, and the goal of the combo isn’t to maximize stimulation but to cover more of the day with a smoother curve. Patients often hit a wall on stim monotherapy in their early thirties when the life-complexity stack jumps, the job demands more than the stim alone covers, the kids and the mortgage want something from the after-work hours, and the four-to-six-hour window of focus that the stim used to be enough for is no longer enough. The patient who’s been “fine on Adderall since college” sometimes comes in at thirty-six saying his focus is falling apart, when the answer is that his life got harder, the dose stayed the same, and the lane is wide enough for a second medication.
What the non-stim class is bad at
The non-stim class is not a substitute for stims in patients where stims actually work. That sentence should be tattooed somewhere visible, because the framing in some corners of the internet and in some prescribers’ offices is that non-stims are the responsible choice and stims are the irresponsible one, and that framing is not honest about the trade-offs the patient is actually making. The non-stim class is slower onset, less reliable in its effect size, harder on the side effect ledger per unit of symptom improvement, and worse at the core attention symptoms in the head-to-head trials. None of that means the class isn’t useful, it means the class is useful in a specific lane and badly chosen outside it.
If a patient is responding well to a stim, switching to a non-stim because of a vague “concern about stims” is a quality-of-life downgrade most of the time. The concern is sometimes worth taking seriously, the patient who finds the stim is making his anxiety worse, the patient whose cardiac history has shifted, the patient whose substance use picture has changed and the controlled substance is now genuinely the wrong tool. But the concern is sometimes a free-floating worry that the patient picked up from a podcast or a relative or a stretch of social media, and in that case the right move is to talk about what the stim is actually doing, what the real risk profile looks like in the patient sitting in the chair, and what he is giving up if he switches. A lot of those conversations end with the patient deciding to stay on the stim he’s responding to, and that decision is the right one when it’s made in plain language with the actual numbers on the table.
Cardiac, BP, HR considerations across the class
The cardiovascular picture is different for each drug in the lane, and the patient deserves to know which way the needle moves for the one he’s starting. Intuniv and Kapvay drop blood pressure by a few points on average and more in some patients, and the orthostatic piece is worth knowing about for the patient who is going to stand up from his desk to take a call. Strattera and Qelbree push blood pressure up by a few points and the heart rate up by a similar small margin, worth tracking with a home cuff in the first three months and at every visit after. Wellbutrin is roughly neutral on the BP picture in most patients, with the occasional outlier. Modafinil is mildly activating on the BP side, smaller than the stim signal but worth noting in patients already running high.
The combo conversation has to factor in the receptor math. Putting Strattera on top of a stim, or Wellbutrin on top of a stim, stacks two BP-elevating drugs and the result is usually a bigger push than either alone, which in some patients is the price of the better symptom control they were after and in others is genuinely a problem. Putting Intuniv on top of a stim is sometimes a useful BP wash, the alpha-2 agonist counteracting some of the stim’s vasoconstriction, but the math is patient-specific. Patients on antihypertensives need careful timing on the alpha-2 agonists especially, dosing at bedtime helps when the BP overlap with the antihypertensive’s peak would otherwise put the patient on the floor. The patient on three antihypertensives already running low-normal BP at home is not the right patient for Intuniv even if his ADHD picture would otherwise fit it, and a Strattera trial or a careful stim trial with home BP tracking is the better lane in that case.
The “I want off stimulants” patient
Some version of this conversation lands in clinic almost every week. The patient is on a stim, has been on it for years, is responding to it, and has decided he wants off. The reasons vary. Sometimes it’s a podcast that scared him about brain shrinkage, sometimes a partner who doesn’t love the personality on the days he forgets the dose, sometimes a vague sense that he’s been on the medication too long. Sometimes it’s a real concern, an emerging cardiac picture, a sleep problem the stim is making worse, a relationship to the medication that has started to feel sticky in a way he doesn’t trust.
The honest framing depends on which version of the conversation it is. If the patient is responding well to his stim and the reasons for switching are mostly external, the right answer is plain words. “You’re responding well to your current medication, the non-stim lane is going to be a downgrade for the core symptoms, but if you want to try it, we can run a six-to-eight week Strattera or Qelbree trial and reassess.” That protocol respects patient autonomy without selling a story that the non-stim is going to magically do what the stim was doing. Most of those trials end with the patient coming back saying the stim was clearly better and asking to go back. He learned what he learned by trying, and now he’s making the choice with actual data instead of the podcast’s data.
If the patient has a real reason to come off the stim, the conversation is different. The non-stim trial gets framed as the right tool for the changed picture, the expectations get set to “this is going to be a different shape of help, not the same shape at lower volume,” and the trial runs eight to twelve weeks with regular check-ins. Sometimes the answer is yes and the patient stays in the new lane. Sometimes the answer is that the original concern about stims was the right read but the non-stim isn’t enough, and the conversation becomes about a different stim molecule or a third path. Either way the patient is doing the choosing in plain language with the actual numbers on the table.
A note on adolescent versus adult use
The pediatric ADHD lane has its own conversation, and the non-stims behave differently in a thirteen-year-old than a thirty-eight-year-old. Strattera, Intuniv, and Kapvay all have bigger pediatric data sets than adult data sets, and the alpha-2 agonists especially are most often picked in kids and teens. The adult lane this article is about leans more on Strattera and Qelbree as monotherapy, Wellbutrin when the picture has the depression or nicotine overlay, modafinil when the stim lane is closed, and the alpha-2 agonists as combo partners with stims more often than as monotherapy.
The work the medication doesn’t do
The pill is leverage, not a substitute for the work. The work, in adult ADHD, is the same work whether the patient is on a stim, a non-stim, both, or neither. Calendar systems that the patient actually uses, not the third app he downloaded and abandoned. Accountability structures, a partner, a coach, a coworker, a written check-in with himself on Sundays, something external that holds the executive function piece when the patient’s own brain isn’t holding it. Sleep schedule, which is the single biggest leverage point most patients have and the one most often ignored, because the lift in ADHD symptoms from going to bed at the same time every night and getting seven and a half hours is bigger than the lift from any non-stim medication in the class. Exercise that builds the dopamine baseline, twenty minutes of cardio four days a week moves the needle on attention and mood in a way that compounds over months. Caffeine timing, the third coffee at 2 p.m. is wrecking the sleep that would have done more for the attention than the third coffee did. Screen-time management at night, the unsexy part of the conversation that nobody wants to have but that every patient who has actually changed it reports back as bigger than the medication’s contribution.
None of this is novel and none of it is glamorous, but it is the actual work and the pill is leverage on the work, not the work itself. A patient on Strattera plus a careful sleep schedule plus four-day-a-week cardio plus a calendar he trusts is in a different life than a patient on Strattera doing none of the rest, and the gap between those two patients is bigger than the gap between Strattera and Adderall. The medication conversation is the easy one, the work conversation is the harder one, and the harder one is where the actual lift comes from.
A few patterns we see often
The afternoon-crash pattern. Patient has been on a long-acting stim for a couple of years, doing fine through the workday, but the 4 p.m. tail of the medication is a cliff. He picks up the kids tired, snaps at his partner over dinner, by 9 p.m. is too wired from the leftover stim to fall asleep but too crashed to do anything useful with the time. Adding a low-dose Intuniv at 3 p.m. smooths the late-day curve and gets him a workable evening with his family. It isn’t dramatic, it is just better, and the better is what compounds.
The ADHD-plus-anxiety pattern. Picture a patient, say, who comes in with attention problems and starts on a methylphenidate product because the prescriber went with the standard first move. Two weeks in, his focus is sharper but his anxiety is at a different level, his chest is tight by mid-morning, his sleep is wrecked, and his partner is asking him what is wrong with him. The underlying anxiety was bigger than anybody appreciated at intake, and the switch is to Strattera. The trial runs the full six weeks because the patient was warned at the start that this is the wait the drug requires. At week seven he is reporting steady focus without the wired edge, and at the three-month mark he is in a better place than the methylphenidate trial ever got him. The first move being a stim wasn’t a failure, it was the standard move that produced the information the patient and the prescriber needed.
The wants-off-stim-but-was-responding pattern. Patient is on Vyvanse, has been for three years, is responding well, and decides he wants off because his new partner doesn’t love the medication conversation. He hears the trade-off in plain language, opts to try Strattera, and runs the six-week trial. At week eight he reports the Strattera is “fine, I guess,” but he can tell the stim was clearly doing more, the morning energy isn’t there, the focus is softer, the productivity at work has dropped by something he can feel. He goes back on Vyvanse, which is the right call, and the partner conversation moves to a different topic because the medication was never what the partner was actually worried about. The trial wasn’t a waste, it gave the patient data instead of speculation.
The pill is leverage on the work. The work belongs to the patient. The non-stim conversation makes that math more obvious, not less.
Where I land on this class
The non-stim lane is real, useful, and underused in the patients it actually fits, and overused as a “responsible alternative” in patients who were already responding to a stim. Both failures are the same failure in slightly different fonts, the failure to match the tool to the picture. Strattera and Qelbree are the monotherapy workhorses, Intuniv and Kapvay are the combo partners that earn their keep with the wound-tight picture and the late-day support, Wellbutrin is the right pick in the ADHD-plus-depression or ADHD-plus-nicotine picture, modafinil and armodafinil are the cards to play when the stim lane is closed for real reasons.
The choices about which drug to try and how to sequence them are yours, the work alongside the medication is yours, and the chemistry is leverage either way, smaller leverage in this lane than in the stim lane but real leverage in the patients where the fit is right. If we’re being honest, the field gets this class wrong in both directions, prescribing it to patients who would have done better on a stim and not offering it to patients who would have done better on a non-stim. Picking the right lane is what the visit is for, the prescription is the end of the conversation, not the start.
Sources
American Academy of Pediatrics Clinical Practice Guideline for the Diagnosis, Evaluation, and Treatment of Attention-Deficit/Hyperactivity Disorder in Children and Adolescents (2019). American Academy of Child and Adolescent Psychiatry Practice Parameter for the Assessment and Treatment of Children and Adolescents with ADHD. FDA prescribing information for atomoxetine (Strattera), viloxazine (Qelbree), guanfacine ER (Intuniv), clonidine ER (Kapvay), bupropion (Wellbutrin), modafinil (Provigil), and armodafinil (Nuvigil). MTA Cooperative Group, multimodal treatment of ADHD long-term follow-up reports. Cochrane Database of Systematic Reviews: Verbeeck W, Bekkering GE, Van den Noortgate W, Kramers C, “Bupropion for attention deficit hyperactivity disorder (ADHD) in adults,” 2017 (CD009504), the directly on-point adult review here, which found low-quality evidence of a benefit; and Kallapiran GK et al, “Atomoxetine for Attention Deficit Hyperactivity Disorder (ADHD) in children and adolescents,” 2012 (CD009804), noting that the atomoxetine Cochrane evidence base is pediatric rather than adult. Selected reports from the Journal of Attention Disorders and the Journal of the American Academy of Child and Adolescent Psychiatry on non-stimulant pharmacotherapy in adults.