Reviewed using peer-reviewed research including the Cochrane review by Hoffman et al. (2021), the meta-analysis by Farhat et al. (2020) in Depression & Anxiety, and randomized trials published in Archives of General Psychiatry, the American Journal of Psychiatry and the Journal of Clinical Psychiatry.
Six compounds have real randomized evidence in trichotillomania: N-acetylcysteine, memantine, clomipramine, olanzapine, naltrexone, and the SSRIs. None of them is approved by the FDA, or by any other regulator, for hair pulling. Every prescription written for this condition is off-label.
The other thing to know before any of the individual numbers make sense: when Cochrane pooled the entire adult drug literature in 2021, it came to ten trials and 286 adults, averaging 29 participants per trial and running between five and thirteen weeks. That is the whole evidence base. A single mid-sized depression trial enrols more people than every adult hair-pulling drug trial in history combined.
So the useful question isn't which drug is best. It's how much confidence each number deserves, and what to ask for when you sit down with a prescriber. If you're still working out what trichotillomania is and how it's diagnosed, start there instead.
How much evidence actually exists
The 2021 Cochrane review of pharmacotherapy for trichotillomania graded each medication class on both effect and certainty. Its conclusion was unambiguous: there was "insufficient evidence from meta-analysis to confirm or refute the efficacy of any agent or class of medication for the treatment of TTM in adults, children or adolescents." Preliminary signals existed for three drugs. Nothing more.
Here is every randomized adult comparison, with the number of people it rests on.
| Compound | Response vs placebo | People in the trial | Certainty grade |
|---|---|---|---|
| Olanzapine | 85% vs 17% (12 wks) | 25 | Low |
| N-acetylcysteine | 56% vs 16% (12 wks) | 50 | Moderate |
| Inositol | 42% vs 32% (10 wks) | 38 | Low |
| Clomipramine | 40% vs 0% (9 wks) | 16 | Very low |
| Naltrexone | 38% vs 25% (6–8 wks) | 68 (2 trials) | Very low |
| Silymarin | 36% vs 29% (6 wks) | 36 | Low |
| SSRIs | No response data extractable | 78 (2 fluoxetine trials) | Not graded |
Memantine isn't in the table because its trial was published in 2023, after the review closed. It's covered below on its own terms.
N-acetylcysteine (NAC)
Typical trial dose: 1,200 to 2,400 mg a day, usually split into two doses.
NAC is an amino acid derivative sold over the counter as a supplement. It acts on glutamate signalling in the nucleus accumbens, a system involved in habit and compulsion rather than mood, which is part of why it behaves differently from antidepressants here.
The detail that gets lost is the timing. Significant improvement did not appear until after nine weeks of treatment. People routinely try NAC for a month, conclude it does nothing, and stop, which is roughly the equivalent of abandoning an antibiotic on day two.
The paediatric picture is different, and parents should know this before spending money. In a separate randomized trial of 39 children and adolescents aged 8 to 17, NAC produced no significant difference from placebo on any primary or secondary measure: 25% were judged responders against 21% on placebo. The authors' own conclusion was that the finding "highlights the importance of referring children with TTM to appropriate behavioral therapy before initiating pharmacological interventions."
Practical notes from the 2022 literature review on NAC in body-focused repetitive behaviors: it costs around ten US dollars for a hundred 600 mg capsules, side effects are dose-dependent, and it warrants caution in people with asthma or a tendency to anaphylaxis, and in people susceptible to fluid overload such as those with congestive heart failure. It also interacts with some medications, which is a pharmacist conversation as much as a doctor one.
SkinAware, the app behind this site, lets you log urges and episodes for free and start a habit reversal course.
Get the appMemantine
Trial dose: 10 to 20 mg a day.
Memantine is licensed for Alzheimer's disease and works on the same glutamate system as NAC by a different route. It produced the largest result in this field to date.
Three caveats belong next to that number. The sample was mixed, combining hair pulling and skin picking rather than testing trichotillomania alone. It is one trial, from one center, and it has not yet been replicated. And the placebo response of 8.3% is unusually low for this literature, which tends to widen the apparent gap.
None of that makes the result less interesting. It makes it a single strong finding rather than a settled fact.
SSRIs
Typical doses: standard antidepressant ranges for fluoxetine, sertraline and the rest.
SSRIs are the medication most people with trichotillomania are actually offered, and they have the weakest case of anything on this page. Cochrane could not extract a single usable treatment-response figure for them from the randomized literature. A separate meta-analysis of behavior therapy against serotonin reuptake inhibitors found a pooled effect size of 0.41 for the drug arm and 1.41 for behavior therapy.
That doesn't mean an SSRI in your medicine cabinet is a mistake. In a well-characterised sample of 85 adults with hair pulling disorder, 78.8% had another lifetime psychiatric diagnosis, with depressive and anxiety disorders prominent among them. Treating those in their own right is a good reason to be on an SSRI. Expecting it to reduce the pulling is where people get let down.
Clomipramine
Typical trial dose: 50 mg titrated up to 250 mg a day.
Clomipramine is a tricyclic antidepressant with strong serotonin reuptake action, and it has the longest history here. A 1989 crossover study of thirteen women found it clearly better than desipramine, a noradrenergic tricyclic, which was the first real evidence that the serotonin system was doing something in this condition.
Cochrane's read of the placebo comparison is a nine-week trial in which 40% responded on clomipramine and 0% on placebo, from sixteen people, graded very low certainty with a confidence interval spanning 0.36 to 90.83. Dropouts due to adverse events ran at 30% against 0%.
A drug with a genuine signal, then, a heavy side-effect load, and a head-to-head trial in which the behavioral arm won.
Olanzapine
Typical trial dose: 2.5 mg titrated up to 10 mg a day. Mean dose at endpoint in the trial was 10.8 mg.
Olanzapine is an atypical antipsychotic. The single randomized trial, run between 2001 and 2005 and published in 2010, was twelve weeks and flexible-dose. Eleven of thirteen participants on olanzapine (85%) were responders against two of twelve on placebo (17%).
It is the biggest separation from placebo in the table and the thinnest evidence underneath it. Twenty-one of the twenty-five participants, 84%, reported at least one adverse event, though none withdrew because of one. Cochrane's summary of the tolerability picture across the whole field was that glutamate modulators had the least severe side-effect profile in adults, while antipsychotics were associated with several adverse effects.
Olanzapine is not a first conversation. It appears in clinical practice when other things have been tried and the pulling is severe.
Naltrexone
Trial dose: 50 mg a day titrated up to 150 mg a day over four weeks.
Naltrexone blocks opioid receptors, and the theory behind using it is that it might blunt the reward some people get from pulling. The theory is more appealing than the data.
An eight-week double-blind trial randomized 51 adults with trichotillomania who reported urges to pull, 25 to naltrexone and 26 to placebo. Naltrexone failed to produce significantly greater reductions in hair pulling than placebo. Cognitive flexibility improved on a lab measure. Participants with a family history of addiction showed a numerically larger drop in urges that did not reach significance, which is a lead for future research rather than a result.
Cochrane pooled two naltrexone trials, 68 people, and found 37.5% response against 25% on placebo, graded very low certainty.
What the trials recorded about side effects
This table reports what the trials themselves logged, which is not the same as a full side-effect profile. Prescribing information covers rare and long-term effects that eight-week studies are too small and too short to detect.
| Compound | What the trials recorded |
|---|---|
| N-acetylcysteine | No adverse events in the adult trial. In the paediatric trial, dropout from adverse events was 5% vs 0% on placebo. Outside trials: nausea, vomiting, diarrhoea and constipation below 2,400 mg/day; fever, chills, rash and headache above it. |
| Memantine | Adverse events did not differ significantly between memantine and placebo. |
| SSRIs | Dropout due to adverse events 5.1% vs 0% on placebo across two fluoxetine trials. |
| Clomipramine | 40% of the group in the CBT comparison trial withdrew early because of side effects. Cochrane recorded 30% vs 0% dropout from adverse events. |
| Olanzapine | 84% of participants reported at least one adverse event. None withdrew because of one. Antipsychotics carried the heaviest adverse-effect burden of any class in the review. |
| Naltrexone | Cochrane found no data on dropouts due to adverse events. The trial itself recorded mild side effects that did not differ from placebo, apart from sedation, which was more frequent on naltrexone. |
How medication compares with the behavioral work
That ordering is stable across every review of this field, including the paediatric one, where the authors of the null NAC trial argued explicitly for behavior therapy before medication in children.
It doesn't mean medication is pointless. It means medication is the second lever, and that starting it instead of the behavioral work rather than alongside it is the pattern most likely to disappoint. Habit reversal training is a specific, learnable protocol, not a vague instruction to be more mindful, and most of what determines whether it works is how well the technique is matched to the kind of pulling you actually do.
Going into the appointment
Prescribing decisions belong to your prescriber. What you can control is the quality of the information you walk in with, and most of these appointments fail on data rather than on willingness.
Four things make a difference.
Ask what response would look like, and by when. For NAC the honest answer is around week nine to twelve, not week four. Agreeing the review point in advance is the best defense against abandoning something too early or staying on something too long.
Say what else you're taking, including supplements. NAC is sold as a supplement, which is exactly why people forget to mention it. It interacts with some medications.
Name any timing pattern. If pulling started or stepped up within weeks of beginning a medication or raising a dose, say so with dates. That's specific, answerable information, and it comes up often enough with stimulants prescribed for ADHD to be worth checking.
If you're a parent, the paediatric evidence is different enough that it's worth stating in the room: the one randomized trial of NAC in children found no benefit, and its authors recommended behavior therapy first.
Frequently asked questions
No. No drug has regulatory approval for hair pulling anywhere. Every prescription is off-label, which is both legal and common, and it means the evidence sits in individual trials rather than in a licence.
Depends what "best" means. N-acetylcysteine has the only adult result Cochrane graded moderate certainty. Memantine produced the largest single result, 60.5% much or very much improved against 8.3% on placebo, but from one unreplicated trial in a mixed hair-pulling and skin-picking sample. Olanzapine has the biggest percentage separation and the smallest sample. None of them outperformed habit reversal training.
Probably not on its own. Cochrane could not extract usable response data for SSRIs, a meta-analysis put serotonin reuptake inhibitors at an effect size of 0.41 against 1.41 for behavior therapy, and the NHS says antidepressants are not usually prescribed for it. They remain a reasonable treatment for co-occurring depression or anxiety, which is a different goal.
Longer than most people wait. In the NAC trial, improvement did not separate from placebo until after nine weeks. Most trials in this field ran between five and thirteen weeks in total. A four-week verdict is premature for almost everything here.
The evidence is thinner and less encouraging. The one randomized paediatric trial, 39 children and adolescents on NAC, found 25% responders against 21% on placebo, no better than chance. Its authors recommended behavior therapy before medication in this age group.
It is sold over the counter in many countries, so physically yes. That doesn't make it free of consequences: it interacts with some medications and warrants caution in asthma and in conditions where fluid overload is a risk. Tell whoever manages your care that you're taking it.
That's the common outcome rather than a failure. Response in these trials means much or very much improved, not stopped. Medication reducing the pressure while behavioral work changes the behavior is how most people who do well actually do well.
What to do next
If you're heading into an appointment, the most valuable thing you can bring is a real record: episodes, sites, urges you resisted, and what was happening around them. That record is also what makes any medication trial measurable afterwards, since "a bit better" is not an outcome anyone can act on.
Logging that in SkinAware takes a few taps per entry, and it tracks resisted urges as well as pulls, which is often where the first change shows up. It also walks through habit reversal training module by module, which is the part of this with the strongest evidence behind it, and there's a moderated community of people working on the same thing.
Get support for picking, pulling and biting
SkinAware helps you log episodes, identify patterns, and see real progress over time.
Available on iOS, Android and in your browser.
Medication decisions are your prescriber's. The data you bring to that conversation is yours.
Continue Reading
References
- Hoffman, J., Williams, T., Rothbart, R., Ipser, J. C., Fineberg, N., Chamberlain, S. R., & Stein, D. J. (2021). Pharmacotherapy for trichotillomania. Cochrane Database of Systematic Reviews, 9, CD007662.
- Farhat, L. C., Olfson, E., Nasir, M., Levine, J. L. S., Li, F., Miguel, E. C., & Bloch, M. H. (2020). Pharmacological and behavioral treatment for trichotillomania: An updated systematic review with meta-analysis. Depression and Anxiety, 37(8), 715–727.
- Grant, J. E., Odlaug, B. L., & Kim, S. W. (2009). N-acetylcysteine, a glutamate modulator, in the treatment of trichotillomania: A double-blind, placebo-controlled study. Archives of General Psychiatry, 66(7), 756–763.
- Grant, J. E., Chesivoir, E., Valle, S., Ehsan, D., & Chamberlain, S. R. (2023). Double-blind placebo-controlled study of memantine in trichotillomania and skin-picking disorder. American Journal of Psychiatry, 180(5), 348–356.
- Bloch, M. H., Panza, K. E., Grant, J. E., Pittenger, C., & Leckman, J. F. (2013). N-acetylcysteine in the treatment of pediatric trichotillomania: A randomized, double-blind, placebo-controlled add-on trial. Journal of the American Academy of Child & Adolescent Psychiatry, 52(3), 231–240.
- Ninan, P. T., Rothbaum, B. O., Marsteller, F. A., Knight, B. T., & Eccard, M. B. (2000). A placebo-controlled trial of cognitive-behavioral therapy and clomipramine in trichotillomania. Journal of Clinical Psychiatry, 61(1), 47–50.
- Van Ameringen, M., Mancini, C., Patterson, B., Bennett, M., & Oakman, J. (2010). A randomized, double-blind, placebo-controlled trial of olanzapine in the treatment of trichotillomania. Journal of Clinical Psychiatry, 71(10), 1336–1343.
- Grant, J. E., Odlaug, B. L., Schreiber, L. R. N., & Kim, S. W. (2014). The opiate antagonist, naltrexone, in the treatment of trichotillomania: Results of a double-blind, placebo-controlled study. Journal of Clinical Psychopharmacology, 34(1), 134–138.
- McGuire, J. F., Ung, D., Selles, R. R., Rahman, O., Lewin, A. B., Murphy, T. K., & Storch, E. A. (2014). Treating trichotillomania: A meta-analysis of treatment effects and moderators for behavior therapy and serotonin reuptake inhibitors. Journal of Psychiatric Research, 58, 76–83.
- Houghton, D. C., Maas, J., Twohig, M. P., Saunders, S. M., Compton, S. N., Neal-Barnett, A. M., Franklin, M. E., & Woods, D. W. (2016). Comorbidity and quality of life in adults with hair pulling disorder. Psychiatry Research, 239, 12–19.
- Lee, D. K., & Lipner, S. R. (2022). The potential of N-acetylcysteine for treatment of trichotillomania, excoriation disorder, onychophagia, and onychotillomania: An updated literature review. International Journal of Environmental Research and Public Health, 19(11), 6370.
- Swedo, S. E., Leonard, H. L., Rapoport, J. L., Lenane, M. C., Goldberger, E. L., & Cheslow, D. L. (1989). A double-blind comparison of clomipramine and desipramine in the treatment of trichotillomania (hair pulling). New England Journal of Medicine, 321(8), 497–501.
- NHS. Trichotillomania (hair pulling disorder). Retrieved July 2026.
Last updated: July 2026