Guanfacine

Guanfacine was approved in the United States as an antihypertensive in 1986 [1], and its extended-release form was approved for ADHD in 2009 as Intuniv [9]. It is now approved for ADHD in the United States, Canada, Europe and Japan, and in Japan for adults too [2][9]. The ADHD use grew out of Arnsten's monkey research on the prefrontal cortex [9].

ADHD in children and adolescents. In fixed-dose trials, 2, 3 and 4 mg a day each lowered ADHD Rating Scale scores by about 16–19 points against 8.5 on placebo [10], with effect sizes of 0.43–0.62 [11]. Added to a stimulant in 461 children with a partial response, it improved symptoms further [12]. In adolescents, flexible doses of 1–7 mg improved symptoms (effect size 0.52) but not school or family functioning [13], and in a randomized-withdrawal trial 49.3% of responders kept on guanfacine relapsed over 26 weeks against 64.9% switched to placebo [14]. Pooling nine monotherapy trials, alpha-2 agonists reduced symptoms with a standardized mean difference of −0.59, and as add-ons −0.36 [3]; in a 2023 meta-analysis of 12 guanfacine trials, 58.5% responded against 29.4% on placebo in trials under 10 weeks [15]. The 2018 network meta-analysis put guanfacine's effect on clinician-rated symptoms at −0.67, similar to methylphenidate, but found no significant effect on teachers' or parents' ratings and more dropouts for side effects than placebo [16].

Tics and autism. In children with both tic disorders and ADHD, guanfacine improved teacher-rated ADHD scores by 37% against 8% and tic severity by 31% against none [17], but in a later trial in children with tics alone it was no better than placebo [18]. In 62 children with autism and hyperactivity, the extended-release form cut hyperactivity scores by 43.6% against 13.2%, with no change on a brief working-memory battery [19].

Adults with ADHD. The 2018 network meta-analysis found no adult trials [16]. In 201 Japanese adults, 4–6 mg a day improved symptoms with an effect size of 0.52, but 19.8% stopped for side effects against 3.0% on placebo [20], and the gains held in a 50-week open extension [2]. A 2023 meta-analysis of two adult trials found an effect size of −0.66 [21], whereas a 2025 network meta-analysis of adult treatments found guanfacine markedly less acceptable than placebo (odds ratio of stopping 3.70) and counted only stimulants and atomoxetine as effective on both self- and clinician-rated symptoms [22].

Cognition in healthy people. In aged monkeys, low doses improved working memory without lowering blood pressure or causing sedation [6]. In healthy young volunteers, a single 29 µg/kg dose improved paired-associates learning but not short-term recognition memory [23], and unlike clonidine it did not slow responses on a demanding attention test [24]. In 60 healthy young men, 1 or 2 mg had no effect on any memory or executive-function test, with trends toward worse backward digit span and slower reactions [7], and in 123 healthy adults over 75, 0.1 or 0.5 mg a day for 12 weeks did not improve executive function, quality of life or global function [8]. In an early comparison in healthy young men, 2–4 mg lowered systolic blood pressure without changing objective performance, its side effects such as tiredness came later and were milder than clonidine's, and 1 mg at night left REM sleep unchanged where clonidine reduced it [25]. Arnsten's group suggests that healthy prefrontal circuits may already work near their ceiling [9]. One month after mild traumatic brain injury, a 2 mg dose improved working memory in injured people but not controls [26].

Stress, craving and other uses. In 33 smokers, 3 mg a day removed the effect of stress on how long they could resist smoking in the laboratory and reduced cigarettes per day during a quit attempt, though no one stayed fully abstinent [27]. In 29 people abstinent from cocaine, up to 3 mg a day lowered cue-induced craving, anxiety and arousal [28]. In veterans with chronic PTSD, two placebo-controlled trials found no benefit [29][30]. Reports of guanfacine for delirium in hospital are case reports and retrospective cohorts, with no randomized trial [31]; Arnsten's group notes that it is being studied in large delirium trials and in open-label studies of cognitive symptoms after COVID-19 [32].

Hypertension. Added to a diuretic, 1 mg at bedtime lowered blood pressure by 14/13 mmHg, while 2–3 mg added side effects without further benefit [33]. Against clonidine it controlled blood pressure about as often (55% against 59% of patients) with less somnolence (21% against 35%) [34].

Alpha-2A selectivity. Guanfacine binds the alpha-2A receptor with 15–20 times the affinity it has for the alpha-2B and alpha-2C subtypes [1]. Clonidine is less selective; Jäkälä's group attributed clonidine's worse effects on attention to its lower alpha-2A/2C selectivity and its affinity for alpha-1 receptors [24]. In aged monkeys the two drugs had opposite dose–response profiles: low doses of guanfacine improved working memory without hypotension or sedation, which appeared only at higher doses, whereas low doses of clonidine impaired memory [6].

Prefrontal working-memory circuits. In the dorsolateral prefrontal cortex, information is held in mind by networks of neurons that keep firing through a delay. Alpha-2A receptors sit on the dendritic spines of these neurons next to HCN channels. Stimulating the receptors lowers cAMP, closing the channels and strengthening the connections between neurons; guanfacine applied to single prefrontal neurons enhanced their delay-period firing, and blocking or knocking down HCN1 channels in the prefrontal cortex improved working memory [5][9]. Guanfacine improved working memory in mice but not in mice whose alpha-2A receptors cannot bind agonists, and infusions into the prefrontal cortex of monkeys and aged rats improved working memory [9]. Arnsten's group argues that these effects are larger in primates because the prefrontal cortex is more developed [9].

Blood pressure. Acting on central alpha-2A receptors, it reduces sympathetic outflow from the brainstem, lowering vascular resistance and heart rate [1]. In hypertensive patients, plasma and urinary norepinephrine and plasma renin activity fell in step with blood pressure [35].

Stress and inflammation. Arnsten's reviews add actions outside the prefrontal cortex: weakening amygdala plasticity, reducing norepinephrine release from the locus coeruleus, and deactivating microglia [9][32]. These are proposed mechanisms for its use in stress-related and neuroinflammatory disorders, drawn mostly from animal work.

The prescribing information states that how guanfacine works in ADHD is not known and that it is not a stimulant [1].

Immediate release against Intuniv. In adults, 1 mg of Intuniv gave a 60% lower peak (1.0 against 2.5 ng/mL), 43% less total exposure, a peak 3 hours later and a similar half-life (18 against 16 hours) compared with 1 mg of immediate release; the label therefore says not to switch milligram for milligram [1]. In five adult smokers, 3 mg a day of immediate release and 4 mg a day of extended release gave the same trough level [36].

Food and handling. A high-fat breakfast raised Intuniv's peak by about 75% and exposure by about 40%, and crushing, chewing or breaking the tablet speeds release [1].

Age and weight. Children had higher levels than adolescents on the same dose (peaks of 10.1 against 7.0 ng/mL at 4 mg a day), which the authors attributed to body weight [37]; the label therefore targets a weight-based range [1].

Sedation. In the fixed-dose pediatric trials, somnolence (including sedation) was reported by 28–51% on 1–4 mg a day against 11% on placebo, rising with dose, along with fatigue (14% against 3%) and hypotension (7% against 3%) [1]. Across pediatric trials of alpha-2 agonists, one extra child had somnolence for every four treated as monotherapy [3]. In a two-year open study in 214 children, 36% reported somnolence, sedative effects peaked in week 3, and 3.3% stopped because of side effects [38].

Blood pressure and heart rate. In a 2025 network meta-analysis, stimulants, atomoxetine and viloxazine raised blood pressure and pulse, whereas guanfacine lowered them, by about 3/2 mmHg and 4 beats per minute in children and adolescents and by about 10/8 mmHg and 7 beats per minute in adults (mostly low or very low confidence) [4]. In the Japanese adult extension, 20% had a fall in blood pressure and 17% bradycardia [2]. The label warns of hypotension, bradycardia, fainting and worsening of heart-block or sinus-node disease, especially with other drugs that slow the heart [1].

QT interval. A meta-analysis flagged QTc prolongation with the extended-release form in pediatric trials [3], but a thorough QT study in 83 healthy adults found no clinically relevant prolongation at 4 or 8 mg of immediate release, after individual heart-rate correction [43].

Withdrawal. The label reports clinically significant, persistent rebound hypertension, and hypertensive encephalopathy, after abrupt discontinuation, often of high doses alongside a stimulant [1]. In 1980s hypertension trials, stopping clonidine abruptly raised blood pressure rapidly whereas stopping guanfacine raised it more gradually [34], and in five patients blood pressure returned to pre-treatment levels over four days without overshoot [35].

Overdose. Overdose causes drowsiness, lethargy, bradycardia and hypotension, sometimes after an initial rise in blood pressure, and the label advises observing children for up to 24 hours because hypotension can be delayed [1]. A 16-year-old who took 25 mg had early hypertension, then orthostatic hypotension and a QTc of 593 ms 30 hours later, and recovered with fluids [44].

Pregnancy. Decades of use have not identified a risk of birth defects, but use in pregnancy has been infrequent [1].

Strong and moderate CYP3A4 inhibitors such as ketoconazole and fluconazole raise guanfacine levels, and inducers such as rifampin and efavirenz lower them; the label halves or doubles the dose accordingly [1]. Antihypertensives, other sympatholytic drugs and drugs that slow the heart add to its effects on blood pressure and conduction, and alcohol and other sedatives add to its sedation [1]. It did not change exposure to methylphenidate or lisdexamfetamine [1], and it has been studied added to long-acting stimulants without new safety signals [12]; rebound hypertension after stopping it has been reported in people also taking a stimulant [1].