In 2009, psychiatrist Jon Grant and two colleagues gave fifty adults with trichotillomania, the compulsive pulling out of one's own hair, twelve weeks of either a placebo or N-acetylcysteine, the same amino acid compound sold in most supplement aisles. Doses ran from 1,200 to 2,400 milligrams a day. By the end, 56 percent of the people taking NAC rated as much or very much improved, against 16 percent of those on placebo, according to the trial published in the Archives of General Psychiatry.

Four years later, a trial of 39 children and adolescents aged 8 to 17 with the same condition reported no significant difference between NAC and placebo on any primary or secondary outcome. Grant was a coauthor on that paper too. The pattern those two results form, one striking and one flat, has repeated across nearly two decades of research on NAC for mental health.

The short answer for anyone weighing the supplement is that NAC has a credible mechanism and a small set of positive randomized trials in compulsive and addictive behaviors, while several of its larger or later trials failed to reproduce those wins. What follows is the record trial by trial, with the sample sizes attached, so the headline numbers can be read in context.

Why Psychiatrists Started Studying a Hospital Antidote

Most people who encounter NAC in a hospital meet it as the standard antidote for acetaminophen overdose or as an inhaled agent that thins mucus. Its appeal to brain researchers came from a different property. NAC supplies cysteine, and cysteine feeds both the body's glutathione supply and a transporter in brain cells that trades cystine in for glutamate out, a system our earlier explainer on what NAC does in the body walks through in more detail.

The study that turned that transporter into a psychiatric target came from a team that included neuroscientist Peter Kalivas and appeared in Nature Neuroscience in 2003. Working with rats, David Baker and colleagues found that withdrawal from repeated cocaine reduced extracellular glutamate in the nucleus accumbens by decreasing the exchange of cystine for glutamate. When the researchers stimulated that exchange with NAC, cocaine-seeking reinstatement was prevented. A rodent result is a long way from a person with a craving, but it gave clinicians a specific reason to try the compound in people whose symptoms looked like loops of urge and relief.

Glutamate is also the reason the same compound shows up in schizophrenia and mood research. Researchers who think of compulsions, cravings and some negative symptoms as problems of glutamate modulation have very few tools that act on that system, which helps explain why an inexpensive, well-tolerated supplement drew so many trials. Speaking about a different glutamate drug in 2023, Grant put the shortage plainly: "We don't have a lot of options about what might modulate glutamate in the brain."

The Trials That Worked

The adult hair-pulling study was the first widely cited win. The second came in 2012, when Kevin Gray and colleagues randomized 116 people aged 15 to 21 with cannabis dependence to 1,200 milligrams of NAC twice a day or a placebo for eight weeks. Participants on NAC had more than twice the odds of producing a negative urine cannabinoid test during treatment, according to the paper in the American Journal of Psychiatry, an odds ratio of 2.4.

Gray described the finding in a National Institute on Drug Abuse podcast. "We found those that received NAC actually had more than twice the likelihood of achieving abstinence during treatment," he said, in remarks recorded before his group's larger adult study had reported.

The schizophrenia result belongs on this side of the ledger as well. In a 24-week trial published in Biological Psychiatry in 2008, Michael Berk's group randomized 140 people with chronic schizophrenia to 1 gram of NAC twice daily or placebo on top of their usual medication, and 84 completed the study. Negative symptoms on the PANSS scale improved by a mean difference of 1.83 points in favor of NAC, with a p value of .018, and akathisia, the restless movement side effect of antipsychotics, also improved. The effect was modest, and the dropout rate was high, which is common in trials of that length.

The Trials That Did Not

The pediatric hair-pulling trial was the first clear miss. The more consequential one arrived in 2017. Gray's team, working through the National Drug Abuse Treatment Clinical Trials Network, ran the adult version of its cannabis study with 302 participants aged 18 to 50, the same 1,200 milligrams twice daily, twelve weeks of treatment and contingency management for everyone. The odds ratio for abstinence came back at 1.00, with a confidence interval from 0.63 to 1.59 and a p value of 0.984, published in Drug and Alcohol Dependence. Same investigator, same dose, a larger sample and an older population, and no detectable effect.

Bipolar disorder followed a similar arc. Berk's group reported benefit for depressive symptoms in bipolar disorder in a 2008 trial. It then ran a maintenance study, published in BMC Medicine in 2012, in which 149 people took 2 grams a day in an eight-week open-label phase before being randomized to continue NAC or switch to placebo. The authors reported no significant between-group differences in recurrence or symptomatic outcomes.

Depression more broadly has not held up under pooling. Reviewing the meta-analyses in the Journal of Clinical Psychiatry in 2021, psychiatrist Chittaranjan Andrade noted that the largest, by Kishi and colleagues, combined seven randomized trials and 728 patients and found NAC was not superior to placebo on depression ratings. A smaller pooling by Nery and colleagues, six trials and 248 patients, found a moderate standardized effect of 0.45. Andrade's conclusion was direct: "At present, therefore, evidence does not encourage the use of NAC as an augmentation treatment for patients with MDD or BD."

Obsessive-compulsive disorder sits in between. A 2024 meta-analysis in Frontiers in Psychiatry pooled six randomized trials covering 195 patients who took NAC alongside an SSRI. The only signal appeared on the total Y-BOCS score at weeks five through eight, at a p value of 0.05, with no differences at other time points or on the subscales. The authors wrote that NAC augmentation may benefit patients with moderate to severe OCD, a carefully hedged sentence resting on fewer than two hundred people.

Reading the Pattern

Lined up by size, the positive results come mostly from single trials of 50 to 140 people in narrowly defined groups, while the larger or later studies in broader populations tend to land on zero. The cannabis pair is the cleanest illustration, because the design, dose and lead investigator stayed constant while the age group and sample size changed. Adolescents in 2012 responded; adults in 2017 did not, even with more than two and a half times as many participants.

Age may matter, and so may the severity of illness, the outcome chosen and whether NAC was added to an existing medication. None of those explanations has been confirmed in a trial built to test it, and readers who see a single study cited as proof that NAC treats anxiety, compulsions or addiction are seeing one row of a longer table. The NAC clinical trial record is a set of leads, several of them promising, few of them replicated.

What This Means on the Shelf

NAC remains legal to sell as a dietary supplement in the United States because of a policy decision rather than a settled definition. In August 2022 the Food and Drug Administration issued final guidance stating its intent to exercise enforcement discretion over the sale of certain NAC products, after a dispute over whether a compound first approved as a drug could be marketed as a supplement. That guidance concerns how the product may be sold, and it makes no finding about psychiatric benefit.

The doses in these trials ran from 1,200 milligrams to 2,400 milligrams a day, delivered under monitoring, often as an add-on to existing treatment. People who take NAC for its antioxidant role, a question our comparison of oral glutathione, NAC and GlyNAC covers, are working from a different body of evidence than someone hoping it will quiet a compulsion or a craving.

For anyone in that second group, the trial record argues for a conversation with a prescriber before a purchase, particularly for people already taking psychiatric medication or living with a condition that has its own established treatments. The compound is inexpensive and generally well tolerated in the studies above, and that profile is exactly why researchers keep returning to it. The 2009 hair-pulling trial is now more than seventeen years old, and the replication that would turn its 56 percent into a standard of care has not yet been published.