buspirone (BuSpar)

no boxed warning, and the two warnings it does carry

the buspirone label has no boxed warning. its warnings section has two items. the first is the maoi warning: "there have been reports of the occurrence of elevated blood pressure when buspirone hydrochloride has been added to a regimen including an maoi", so the label recommends against using them together, and the contraindications section makes it absolute for maois taken for depression, within 14 days either way, because of "an increased risk of serotonin syndrome and/or elevated blood pressure"; starting buspirone in anyone on the antibiotic linezolid or intravenous methylene blue is contraindicated for the same reason [label, warnings; contraindications; dosage and administration].

the second is serotonin syndrome, "a potentially life-threatening" condition reported with buspirone alone "but particularly" with other serotonergic drugs, including triptans, with maois, or "with antipsychotics or other dopamine antagonists". the signs: agitation, hallucinations, delirium or coma; racing heart, swinging blood pressure, dizziness, sweating, flushing, fever; tremor, rigidity, muscle jerking, overactive reflexes, incoordination; seizures; nausea, vomiting, diarrhoea. the label says to monitor for it, to stop buspirone and any other serotonergic or antidopaminergic drug immediately if it occurs, that combining buspirone with tryptophan "is not recommended", and that a triptan alongside it calls for careful observation, especially at the start and at dose increases [label, warnings]. the warnings section closes with a sentence that explains why the dopamine point matters: buspirone "has no established antipsychotic activity" and "should not be employed in lieu of appropriate antipsychotic treatment" [label, warnings].

it will not cover benzodiazepine withdrawal

the precautions section states it directly: because buspirone does "not exhibit cross-tolerance with benzodiazepines and other common sedative/hypnotic drugs, it will not block the withdrawal syndrome often seen with cessation of therapy with these drugs". so, before starting buspirone, the label says "it is advisable to withdraw patients gradually, especially patients who have been using a cns-depressant drug chronically, from their prior treatment" [label, precautions]. it describes what that withdrawal can look like: "any combination of irritability, anxiety, agitation, insomnia, tremor, abdominal cramps, muscle cramps, vomiting, sweating, flu-like symptoms without fever, and occasionally, even as seizures", over a period that depends on the drug and its half-life [label, precautions]. the practical meaning is that switching from a benzodiazepine to buspirone is two separate events, a taper and a start, and the taper does not get easier because buspirone is there. the alprazolam explainer on this site has that label’s own taper language [alprazolam-explainer]. the pharmacology section explains the gap: buspirone "has no significant affinity for benzodiazepine receptors and does not affect gaba binding", acts mainly at serotonin 5-ht1a receptors with moderate affinity for dopamine d2 receptors, and, unlike benzodiazepines, has no anticonvulsant or muscle-relaxant effect and "lacks the prominent sedative effect" [label, clinical pharmacology]. its mechanism of action "is unknown" [label, clinical pharmacology].

the interactions: cyp3a4, grapefruit, and the numbers

buspirone is broken down by the liver enzyme cyp3a4, and the label reports what happens when that enzyme is blocked, in healthy volunteers given a single 10 mg dose: verapamil raised exposure 3.4-fold and diltiazem 5.5-fold; erythromycin raised the peak 5-fold and total exposure 6-fold; itraconazole raised the peak 13-fold and total exposure 19-fold; nefazodone raised the peak up to 20-fold and total exposure up to 50-fold; and grapefruit juice, 200 ml double-strength three times a day for two days, raised the peak 4.3-fold and total exposure 9.2-fold [label, drug interactions]. with erythromycin the label recommends 2.5 mg twice a day, with itraconazole or nefazodone 2.5 mg once a day, and with any potent cyp3a4 inhibitor "a low dose of buspirone used cautiously"; it names ketoconazole and ritonavir as others of the kind [label, drug interactions]. the patient information puts it in one line: "avoid drinking large amounts of grapefruit juice" [label, information for patients]. a review of the grapefruit literature, which lists buspirone among the psychiatric drugs affected, explains why the amount and timing matter: the fruit inactivates the gut enzyme irreversibly, so one whole grapefruit or 200 ml of juice is enough and the effect is still half its maximum ten hours later and a quarter at 24 hours, with older people "an especially vulnerable population" [grapefruit].

the opposite also happens: rifampin cut buspirone’s peak by 83.7% and total exposure by 89.6%, and the label names dexamethasone, phenytoin, phenobarbital and carbamazepine as enzyme inducers that may do the same, so the dose "may need adjusting to maintain anxiolytic effect" [label, drug interactions]. food raises buspirone’s bioavailability (total exposure up 84% and peak up 116% in eight volunteers), which is why the label asks for it to be taken "consistently, either always with or always without food" [label, clinical pharmacology; dosage and administration]. haloperidol levels rose with buspirone, nordiazepam rose about 15% with diazepam, cimetidine raised the buspirone peak 40%, one report linked buspirone plus trazodone to three- to sixfold rises in the liver enzyme alt that a second study could not reproduce, and one report of a prolonged prothrombin time on warfarin involved a patient on four other drugs [label, drug interactions]. liver and kidney impairment raised exposure 13-fold and 4-fold, and the label says buspirone "cannot be recommended" in severe impairment of either [label, clinical pharmacology; precautions]. a laboratory note: buspirone can read as metanephrine in a urine test for pheochromocytoma and give a false positive, and the label says to stop it for at least 48 hours before such a collection [label, precautions].

the other precautions, in plain words

this is an old-format label with no numbered warnings list; the items below are the rest of the precautions and population sections.

• driving and alcohol: studies "indicate that buspirone hydrochloride is less sedating than other anxiolytics and that it does not produce significant functional impairment", but its effects "in any individual patient may not be predictable", so the label asks for caution with cars and machinery until a person knows how it affects them. formal studies found it did not add to alcohol’s impairment of performance; the label still calls it "prudent to avoid" the combination [precautions].
• movement disorders: because buspirone binds dopamine receptors, "a question has been raised" about dystonia, pseudo-parkinsonism, akathisia and tardive dyskinesia. controlled trials "failed to identify any significant neuroleptic-like activity", but "a syndrome of restlessness, appearing shortly after initiation of treatment, has been reported in some small fraction" of patients, and the post-marketing list includes cogwheel rigidity, dystonic reactions, parkinsonism, akathisia, acute and tardive dyskinesias and restless legs [precautions; adverse reactions].
• children: two placebo-controlled 6-week trials in 559 patients aged 6 to 17 with generalised anxiety, at 15 to 60 mg a day, found "no significant differences between buspirone and placebo". blood levels in children were equal to or higher than adults’ at the same dose, no unexpected safety findings arose, and "there are no long-term safety or efficacy data in this population" [pediatric use]. the label does not state an age limit in the indication; the trials it reports in children were negative.
• older people: in one study of 6,632 patients, 605 were over 65 and 41 over 75, and their safety and efficacy profiles were similar to younger patients’; age did not change the drug’s blood levels, though "greater sensitivity of some older patients cannot be ruled out" [geriatric use].
• pregnancy and breastfeeding: no harm in rats and rabbits at about 30 times the maximum human dose, but "adequate and well-controlled studies during pregnancy have not been performed", so the label says to use it "only if clearly needed". whether it passes into human milk "is not known" (it does in rats), and the label says to avoid it in nursing women "if clinically possible" [pregnancy; nursing mothers].
• not a controlled substance: "in human and animal studies, buspirone has shown no potential for abuse or diversion and there is no evidence that it causes tolerance, or either physical or psychological dependence". the label still asks prescribers to evaluate for a history of drug abuse and to watch for dose escalation or drug-seeking, because "it is difficult to predict from experiments the extent to which a cns-active drug will be misused" once marketed [drug abuse and dependence].
• overdose: healthy male volunteers were given up to 375 mg a day in pharmacology trials, with nausea, vomiting, dizziness, drowsiness, small pupils and stomach upset as that dose approached. "no deaths have been reported following overdosage with buspirone hydrochloride alone"; the rare fatal cases "were invariably associated with ingestion of multiple drugs and/or alcohol". there is no antidote [overdosage].
• everyday stress: the indications section says "anxiety or tension associated with the stress of everyday life usually does not require treatment with an anxiolytic" [indications].

what the label says about its own trials

the label has no clinical studies section, but it says more about its trials than most old labels do. efficacy "has been demonstrated in controlled clinical trials of outpatients whose diagnosis roughly corresponds to generalized anxiety disorder", defined by the third edition of the psychiatric manual; many of those patients also had depressive symptoms, and their anxiety had lasted from one month to over a year, six months on average [label, indications]. about 2,200 patients took part in the premarketing efficacy trials, which lasted 3 to 4 weeks; the side-effect table pools 17 placebo-controlled 4-week trials; and over 3,500 people were exposed across the whole premarketing programme, about 3,000 of them with safety data systematically collected, mostly at 10 to 30 mg a day [label, adverse reactions]. what the label does not give is a result: no scale, no score, no response rate from any of those trials.

on duration, the label is candid: effectiveness "in long-term use, that is, for more than 3 to 4 weeks, has not been demonstrated in controlled trials. there is no body of evidence available that systematically addresses the appropriate duration of treatment for gad." it adds that in one study of long-term use, 264 patients took buspirone for a year "without ill effect", which is a safety observation, not a measure of whether it kept working, and it asks the prescriber who uses it for extended periods to "periodically reassess the usefulness of the drug for the individual patient" [label, indications]. the dose it describes is 15 mg a day to start, rising by 5 mg a day every two or three days to a maximum of 60 mg, with 20 to 30 mg a day the common range in the trials [label, dosage and administration]. the fda has not published its 1986 review, so the number of pivotal trials and their individual results are not public; the approval record gives the date and the application number [approval].

side effects, with the label’s numbers

in the 17 pooled four-week trials, 477 people on buspirone against 464 on placebo: dizziness 12% against 3%, drowsiness 10% against 9%, nausea 8% against 5%, headache 6% against 3%, nervousness 5% against 1%, fatigue 4% against 4%, insomnia 3% against 3%, dry mouth 3% against 4%, lightheadedness 3% against under 1%, and at 2% against under 1% each, excitement, anger or hostility, confusion, blurred vision, diarrhoea, numbness and weakness [label, adverse reactions, table]. the label’s own summary of what buspirone added over placebo is "dizziness, nausea, headache, nervousness, lightheadedness, and excitement" [label, adverse reactions]. about 10% of the 2,200 patients in the efficacy trials stopped because of an adverse event: 3.4% for central nervous system complaints, mainly dizziness, insomnia, nervousness, drowsiness and lightheadedness; 1.2% for gastrointestinal, mainly nausea; 1.1% for headache and fatigue; and 3.4% for multiple complaints [label, adverse reactions]. the label warns that these figures "cannot be used to predict the incidence of side effects in the course of usual medical practice".

across the wider premarketing database, nonspecific chest pain, disturbed dreams, tinnitus, sore throat and nasal congestion were "frequent" (at least 1 in 100); fainting, low and high blood pressure, depersonalisation, euphoria, akathisia, hallucinations, involuntary movements, suicidal ideation, seizures, raised or lowered libido, raised liver enzymes and weight gain were "infrequent" (1 in 100 to 1 in 1,000); and galactorrhea, delayed ejaculation and impotence were "rare" [label, adverse reactions]. the label notes that without controls in some of those studies "a causal relationship to buspirone hydrochloride treatment cannot be determined". post-marketing reports add allergic reactions and angioedema, serotonin syndrome, the movement disorders listed above, emotional lability, "transient difficulty with recall", urinary retention and tunnel vision [label, adverse reactions]. drowsiness at 10% against 9% is the number most worth noticing beside the benzodiazepine labels on this site: the label’s claim of less sedation is borne out by its own table, and the independent trials below say the same.

what the research says, in both directions

in favour. the cochrane review of azapirones (the class buspirone belongs to) for generalised anxiety pooled 36 randomised trials with 5,908 participants and found they beat placebo, with roughly one extra person improved for every four to five treated (number needed to treat 4.4) [chessick-2006]. the 2019 network meta-analysis of 89 trials in 25,441 people with generalised anxiety found buspirone more effective than placebo and well tolerated; it had no external funding [slee-2019]. a 2023 meta-analysis of 70 randomised trials found azapirones beat placebo for generalised anxiety, about 1.6 times the placebo response rate in a comparison of 2,567 patients, with less sedation than benzodiazepines [rossano-2023]. in the government-funded star*d trial, 565 depressed adults whose ssri had not worked were randomised to add buspirone or bupropion, and about three in ten in each group reached remission [trivedi-2006]; that is an unapproved use, tested for depression rather than anxiety, and the label does not mention it.

against, or smaller than it sounds. the cochrane review found azapirones no better than benzodiazepines, and patients found them less acceptable [chessick-2006]. the network meta-analysis’s finding for buspirone "rested on smaller samples than the evidence for antidepressants such as duloxetine, venlafaxine, and escitalopram" [slee-2019]. the 2023 review found no benefit in panic disorder or social anxiety, which matters because the label’s indication is generalised anxiety and the trials behind it were in that condition; several of its authors disclosed speaker fees and grants from drug companies [rossano-2023]. in a 4-week trial of 244 adults comparing hydroxyzine, buspirone and placebo, buspirone’s advantage over placebo appeared only on secondary measures while hydroxyzine beat placebo on the main one [lader-1998]. in star*d, buspirone was slightly less well tolerated than bupropion [trivedi-2006]. and the label’s own trials in 559 children were negative [label, pediatric use]. nothing on this page runs longer than the label’s one-year safety observation, and nothing measures effectiveness past a few weeks; the label says so. hydroxyzine, the other non-benzodiazepine anxiety drug on this site, has its own explainer [hydroxyzine-explainer].

questions people ask

does buspirone (buspar) have a black box warning?

no. the buspirone label carries no boxed warning. its warnings section has two items: a contraindication with maois, including the antibiotic linezolid and intravenous methylene blue, because of serotonin syndrome and raised blood pressure, and a serotonin syndrome warning that applies to buspirone alone and more so with triptans, antipsychotics and other serotonergic drugs. its precautions cover driving, the fact that it will not block benzodiazepine withdrawal, a small fraction of patients who develop restlessness soon after starting, drug and grapefruit interactions through the liver enzyme cyp3a4, two negative trials in children, and avoidance in severe liver or kidney disease. buspirone is not a controlled substance, and the label reports no evidence of dependence or abuse potential.

is buspirone addictive?

the label says no, in unusually direct language: "in human and animal studies, buspirone has shown no potential for abuse or diversion and there is no evidence that it causes tolerance, or either physical or psychological dependence". in two double-blind studies, volunteers with a history of recreational drug or alcohol use could not tell buspirone from placebo, while they showed a clear preference for methaqualone and diazepam. it is not a controlled substance. the label still asks prescribers to screen for a history of drug abuse and to watch for dose escalation, because how a drug is misused once marketed is hard to predict from experiments. the label has no withdrawal section for buspirone itself.

can buspirone replace a benzodiazepine like xanax or klonopin?

not in the way the question usually means. the label says buspirone does not exhibit cross-tolerance with benzodiazepines and "will not block the withdrawal syndrome often seen with cessation of therapy with these drugs". so a person switching has two separate things happening: a benzodiazepine taper, which the label says should be gradual and ideally completed before buspirone starts, and a new drug that takes time to work. the withdrawal the label describes includes irritability, anxiety, agitation, insomnia, tremor, cramps, vomiting, sweating, flu-like symptoms and occasionally seizures, and none of it is prevented by buspirone. the cochrane review found buspirone-type drugs no more effective than benzodiazepines for generalised anxiety and less acceptable to patients; a 2023 review found them less sedating. the schedule is the prescriber’s.

does buspirone work for anxiety?

for generalised anxiety, yes, modestly, over a few weeks. the cochrane review pooled 36 trials in 5,908 people and found buspirone-type drugs beat placebo, with about one extra person improved for every four to five treated; a 2019 network meta-analysis of 89 trials in 25,441 people found buspirone more effective than placebo and well tolerated, though on smaller samples than the antidepressants it was compared with; a 2023 meta-analysis of 70 trials found about 1.6 times the placebo response rate. the same reviews found no advantage over benzodiazepines, less acceptability to patients, and no benefit in panic disorder or social anxiety. the label’s own trials ran 3 to 4 weeks, and it says effectiveness beyond that "has not been demonstrated in controlled trials". two 6-week trials in 559 children found no difference from placebo.

why can’t you drink grapefruit juice with buspirone?

because grapefruit blocks the gut enzyme cyp3a4 that breaks buspirone down. in the label’s study, 200 ml of double-strength grapefruit juice three times a day for two days raised buspirone’s peak blood level 4.3-fold and total exposure 9.2-fold, and the patient information says to "avoid drinking large amounts of grapefruit juice". a review of the grapefruit literature explains why timing does not solve it: the fruit inactivates the enzyme irreversibly, so one whole grapefruit or 200 ml of juice is enough and the effect is still half its maximum ten hours later. the same enzyme is why erythromycin, itraconazole, nefazodone, verapamil and diltiazem raise buspirone levels 3- to 50-fold and why the label names 2.5 mg doses alongside some of them.

what are the most common buspirone side effects?

in the label’s 17 pooled four-week trials, 477 people on buspirone against 464 on placebo: dizziness 12% vs 3%, drowsiness 10% vs 9%, nausea 8% vs 5%, headache 6% vs 3%, nervousness 5% vs 1%, lightheadedness 3% vs under 1%, and excitement, anger or hostility, confusion, blurred vision, diarrhoea, numbness and weakness each 2% vs under 1%. fatigue, insomnia and dry mouth were no more common than on placebo. about 10% of the 2,200 patients in the efficacy trials stopped because of an adverse event, a third of them for dizziness, insomnia, nervousness, drowsiness or lightheadedness. the label also reports a syndrome of restlessness in a small fraction of patients soon after starting, and post-marketing reports of movement disorders, serotonin syndrome and transient difficulty with recall. the label says drowsiness is usually less than with other anxiolytics, and its own table bears that out.

how long does buspirone take to work, and how long can you take it?

the label does not give a time to effect. what it gives is a dosing schedule, 15 mg a day rising by 5 mg every two or three days to a usual 20 to 30 mg and a maximum of 60 mg, and trials that lasted 3 to 4 weeks. on duration it says effectiveness "for more than 3 to 4 weeks, has not been demonstrated in controlled trials", that "there is no body of evidence available that systematically addresses the appropriate duration of treatment", that 264 patients took it for a year without ill effect, and that the prescriber should reassess its usefulness periodically. the independent trials in the library run a few weeks too. the label has no section on stopping buspirone; whether that silence means stopping is uneventful is not something the label says, and this page does not either.

sources for this section

the library rows link to our one-page summary of each study, with the original paper, its funding and our evidence grade on that page.

1. buspirone hydrochloride tablets prescribing information, fda label on dailymed (set id ed2b56fe-e02f-46bc-808a-352f9bbb5b57, version 3, effective august 27, 2026; remedyrepack, anda 210907): clinical pharmacology, indications and usage, contraindications, warnings, precautions (including information for patients, drug interactions, pregnancy, nursing mothers, pediatric use, geriatric use), adverse reactions with the pooled-trial table, drug abuse and dependence, overdosage, dosage and administration
2. drugs@fda, nda 018731 (buspar, buspirone hydrochloride), approved september 29, 1986
3. ahrq medical expenditure panel survey, prescribed medicines public-use files 2014 (hc-168a) and 2024 (hc-254a): people with at least one buspirone purchase and prescriptions, weighted; the 2014 to 2024 prescription change clears the survey’s margin of error; computed by veisund, method on the usage section of this page
4. chessick ca, et al. azapirones for generalized anxiety disorder. cochrane database of systematic reviews 2006 (36 trials, 5,908 participants) (library row)
5. slee a, et al. pharmacological treatments for generalised anxiety disorder: a systematic review and network meta-analysis. lancet 2019 (89 trials, 25,441 participants) (library row)
6. rossano f, et al. the efficacy, safety, and adverse events of azapirones in anxiety disorders: a systematic review and meta-analysis of randomized controlled trials. european neuropsychopharmacology 2023 (70 trials) (library row)
7. trivedi mh, et al. medication augmentation after the failure of ssris for depression. new england journal of medicine 2006 (star*d, 565 adults; an unapproved use) (library row)
8. lader m, scotto jc. a multicentre double-blind comparison of hydroxyzine, buspirone and placebo in patients with generalized anxiety disorder. psychopharmacology 1998 (244 adults, four weeks) (library row)
9. bailey dg, dresser g, arnold jm. grapefruit-medication interactions: forbidden fruit or avoidable consequences? cmaj 2013 (narrative review; buspirone among the drugs listed) (library row)
10. what the hydroxyzine (vistaril) warnings mean, the companion explainer on this site for the other non-benzodiazepine anxiety drug
11. what the alprazolam (xanax) warnings mean, the companion explainer on this site, with the benzodiazepine label’s own taper language

Most-reported reactions

• Fatigue800
• Nausea798
• Headache731
• Drug ineffective670
• Anxiety668
• Dyspnoea583
• Pain578
• Diarrhoea511
• Depression503
• Dizziness500