does hydroxyzine work for anxiety — what the evidence actually shows

an allergy drug from the 1950s, prescribed for anxiety because it is not a benzodiazepine and not an antidepressant. the label approves it, in language nobody has used since the 1960s, and admits it was never tested past four months.

here is what the record holds. what the label approves and for how long, what the five placebo-controlled trials found when cochrane pooled them, the actual margins in the two largest trials, where the drug sits in the 2019 network of 22 anxiety drugs, the two modern trials in which it did nothing, and the heart-rhythm warning quoted from the regulator that wrote it, with the dose limits it set. eight sources, funding printed where the record states it.

approved for anxiety, in 1950s words, for four months

the us label indicates hydroxyzine “for symptomatic relief of anxiety and tension associated with psychoneurosis and as an adjunct in organic disease states in which anxiety is manifested”, alongside itching from allergic conditions and sedation before and after anaesthesia^[6]. the anxiety dose is “50 to 100 mg q.i.d.”, four times a day^[6]. and then the sentence most prescriptions outlive: “the effectiveness of hydroxyzine as an antianxiety agent for long term use, that is more than 4 months, has not been assessed by systematic clinical studies. the physician should reassess periodically the usefulness of the drug for the individual patient”^[6].

the cochrane review: five trials, 884 people

cochrane searched to march 2010 and found five randomised trials of hydroxyzine in generalised anxiety disorder, against placebo, benzodiazepines or buspirone^[1]. pooled, the trials “provide some evidence that hydroxyzine is more effective than placebo for gad (odds ratio 0.30, 95% ci 0.15 to 0.58)”, with acceptability no different from placebo^[1]. against the active drugs, “hydroxyzine was equivalent in terms of efficacy, acceptability and tolerability” (against chlordiazepoxide odds ratio 0.75, 0.35 to 1.62; against buspirone 0.76, 0.40 to 1.42), and “was associated with a higher rate of sleepiness/drowsiness than the active comparators” (odds ratio 1.74, 0.86 to 3.53)^[1].

the reviewers' verdict, in full: “even though more effective than placebo, due to the high risk of bias of the included studies, the small number of studies and the overall small sample size, it is not possible to recommend hydroxyzine as a reliable first-line treatment in gad”^[1]. no trial has been added to that base that changes it; the review has not been updated since.

the two largest trials, and the margins

the twelve-week trial randomised 334 outpatients with generalised anxiety disorder to hydroxyzine 50 mg a day, bromazepam 6 mg a day, or placebo, after two weeks of placebo run-in^[2]. on the hamilton anxiety scale, hydroxyzine fell 12.16 points and placebo 9.64 (p = .019)^[2]: a margin of about two and a half points on a 56-point scale, over a placebo response that was itself large. response and remission rates also favoured the drug, and “the study showed no statistically significant difference between hydroxyzine and bromazepam. except for drowsiness, which was more frequent with bromazepam, safety results were comparable in the 3 groups”^[2].

the four-week trial randomised 244 primary-care patients in france and the uk to hydroxyzine 50 mg a day, buspirone, or placebo^[3]. on the primary analysis, “a significant difference only between hydroxyzine and placebo” appeared on the hamilton scale: 10.75 points against 7.23^[3]. buspirone separated from placebo only on the secondary measures. neither paper's abstract states who paid for the trial; hydroxyzine's manufacturer is not named on the records we read, and the funding column in the sources below says unknown^[2],[3].

two trials where it did nothing

outside generalised anxiety disorder the modern trials are less kind. in 140 prehospital patients with severe acute pain, adding hydroxyzine to intravenous morphine did not reduce pain and did not reduce anxiety (91% versus 78% with no severe anxiety at fifteen minutes, not significant); the authors concluded it “is not indicated based on our results”^[7]. in 165 children aged 2 to 16 facing outpatient surgery, hydroxyzine “was not effective to control” pre-operative anxiety; the only thing that held anxiety steady was hydroxyzine combined with clowns^[8]. neither is the drug's approved use, and both are single-dose settings. they are here because they are the two most recent placebo-controlled tests of the drug's anxiolytic effect that we could find, and both were negative.

the heart-rhythm warning, from the regulator that wrote it

on 13 february 2015 the european medicines agency's safety committee finished its review of hydroxyzine. it “considered that hydroxyzine was associated with a small but definite risk of qt interval prolongation and torsade de pointes (alterations in the electrical activity of the heart that can lead to abnormal heart rhythms and cardiac arrest)”, and that “the risk did not differ between indications”^[5]. the drug could stay on the market with conditions: “using the medicine at the lowest effective dose for as short a time as possible. use is not recommended in the elderly. the maximum daily dose should be no more than 100 mg in adults (50 mg in the elderly if use cannot be avoided)”, and “use must be avoided in patients who already have risk factors for heart rhythm disturbances or are taking other medicines that increase the risk of qt prolongation”^[5].

the us label carries the same finding as a precaution: “cases of qt prolongation and torsade de pointes have been reported during post-marketing use of hydroxyzine. the majority of reports occurred in patients with other risk factors for qt prolongation/tdp (pre-existing heart disease, electrolyte imbalances or concomitant arrhythmogenic drug use)”, and it contraindicates the drug “in patients with a prolonged qt interval”^[6]. what the us label does not do is set the european dose ceiling or the age restriction: its anxiety dose remains 50 to 100 mg four times a day, four times the eu maximum^[5],[6]. two regulators read the same reports and set different limits.

the rest of the side-effect record

the label reports adverse reactions without percentages. the common ones are anticholinergic dry mouth and drowsiness that “is usually transitory and may disappear in a few days of continued therapy or upon reduction of dose”; patients “should be warned” and “cautioned against driving a car or operating dangerous machinery”, and the effects of alcohol “may be increased”^[6]. post-marketing reports add qt prolongation and torsade de pointes, a rare serious skin reaction (acute generalised exanthematous pustulosis), headache and hallucinations^[6]. for older adults the label says “sedating drugs may cause confusion and over sedation in the elderly; elderly patients generally should be started on low doses of hydroxyzine and observed closely”^[6]. the label contains no dependence or withdrawal section, which is the reason it is often chosen over a benzodiazepine; the cochrane pooled comparison found it no more and no less effective than one^[1].

questions worth taking to whoever prescribed it

• is this for a few weeks or indefinitely? the label says effectiveness past four months was never assessed^[6]
• has anyone checked my qt interval, my potassium, and my other medicines against the list that raises the risk^[5],[6]
• if i am over 65, was the european “not recommended in the elderly” finding weighed^[5]
• what total daily dose am i on, and how does it compare with the 100 mg ceiling one regulator set^[5],[6]
• was cognitive behavioural therapy offered? every anxiety guideline puts it before or beside a drug, and no trial on this page compared hydroxyzine with it^[1]