Pagoclone is a pharmaceutical substance that was under development for stuttering until the 2010s and was previously investigated for panic, anxiety, and premature ejaculation. It does not have a benzodiazepine structure, but it operates at the benzodiazepine site.

Although it had some efficacy in panic and stuttering, it was not sufficient to maintain investment, so the drug has never been commercialized.

It has occasionally been sold online, but its use outside of clinical research settings is minimal.

Pagoclone = RP 59037 (the racemate)

PubChem: 131664

Molecular formula: C23H22ClN3O2

Molecular weight: 407.898 g/mol

IUPAC: 2-(7-chloro-1,8-naphthyridin-2-yl)-3-(5-methyl-2-oxohexyl)-3H-isoindol-1-one


The dosing in medical trials was 0.3 to 0.6 mg/d (oral). The dose was split across two or three administrations per day.

Larger doses upwards of 5 mg have sometimes been used outside of those settings. Those amounts may come with more nonmedical/recreational properties, as seen in an abuse potential study that used 4.8 mg (de Wit, 2006).


When given orally, it seems to last for around 4 to 6 hours.

Experience Reports




It was primarily studied for stuttering and anxiety/panic. In both cases it was found to be effective (Maguire, 2010; Sandford, 2001) but larger trials didn’t demonstrate enough efficacy to justify further development. Pagoclone also did not have adequate efficacy for premature ejaculation and no results were ever published for its use in that condition.


An 8-week controlled trial of pagoclone for stuttering in 132 participants found it significantly improved symptoms while producing minor side effects (Magquire, 2010). Similarly positive results were seen in the 119 individuals who elected to continue treatment in a yearlong open-label extension period. When a larger group of people was studied, the results were not as significant, leading to an abandonment of research even though pagoclone may indeed have some efficacy in the condition.

(Maguire, 2010) – Pagoclone is effective in the treatment of stuttering
  • Background
    • Approximately 1000 subjects have received pagoclone in previous studies of generalized anxiety or panic disorder. A small number of those patients were reported to stutter during screening assessments and it was observed that 0.30 to 0.60 mg/d pagoclone led to a reduction of stuttering into those individuals.
      • 2 cases in the panic disorder trial showed improvement of symptoms related to anxiety and persistent developmental stuttering. Stuttering then returned to baseline after the trial.
  • 8-week, randomized, double-blind study with a 1-year open label extension
  • Pagoclone 0.30 mg was given twice daily.
  • 132 participants: 44 in placebo and 88 in pagoclone
    • After completion of the double-blind phase, 119 subjects entered the open-label extension.
  • Results
    • The pagoclone group had significant benefits for most key variables at Week 4 and Week 8, or at the average of the two time points.
      • There were significant benefits for SSI-3 frequency and duration subscore and for SSS Severity Subscore.
    • Though not significantly different from placebo at Week 8, the pagoclone group clearly had a greater reduction in the percent of syllables stuttered vs. placebo (-7.89 in placebo vs. -17.41 in pagoclone)
    • The average reduction in the % of stuttering events in the pagoclone group was approximately 20%, whereas for placebo it was around 5%.
    • The improvements were still observed through the open-label portion of the study. Patients who completed the open-label phase had a 37.19% reduction in the % of syllables stuttered from baseline at 3 months and a 40.03% reduction at Month 12.
    • Safety
      • The adverse effects were similar between groups. Headache occurred in 6.8% of placebo vs. 12.5% of pagoclone participants. And fatigue occurred in 8% of pagoclone participants but in 0% of placebo patients.
      • No clinically significant changes in any lab, ECG, or vital sign assessments.
  • COI: Study was funded by Indevus Pharmaceuticals. The authors have received money from pharmaceutical companies including Endo Pharmaceuticals, Teva Neuroscience, Indevus Pharmaceuticals, Schering-Plough, Merck, and Eli Lilly.
  • Responses
    • (Ingham, 2010) – Critical response to the (Maguire, 2010) trial
      • There are limitations about the study that were not mentioned and their absence raises questions about the validity of the data. One issue is the use of the Stuttering Severity Instrument, which has questionable reliability and validity in part because it assesses just 200 syllables of speech and the data is collected in the clinical setting.
      • There’s also no evidence the 40% reduction is greater than the natural variability in symptoms. Data on speech performance should have been collected pre-treatment. A review of pharmacological treatments found a minimum reduction of symptoms of 50% should be used since there is such a high level of variability in stutterers.
      • Non-drug treatments, which the authors failed to mention, can produce beneficial changes closer to 90%.
      • Since the data shows the response was no longer significantly better than placebo at 8 weeks and there was no control group for the 12-month data, pagoclone does not seem very effective.
      • One of the individuals collecting stuttering data allegedly had no training in that area.
    • (Maguire, 2010) – A response to (Ingham, 2010) in defense of the trial
      • Based on the open-label portion of the study, improvements continued in the long-term unlike with the behavioral treatments that are available. Further, those behavioral treatments often operate by disrupting the natural flow of speech, which can be disturbing to the speaker, and long-term adherence and treatment effects are poor, in their opinion.
      • They deny any of the people working on the study were not properly trained.


Panic Disorder

A trial of 16 patients with panic disorder with agoraphobia was somewhat positive (Sandford, 2001). Pagoclone at 0.1 mg three times per day significantly reduced the frequency of panic attacks, though without any consistent impact on other measures of anxiety or condition severity. The impact, while statistically significant, only involved a reduction in panic attacks by 2.2 per two-week period vs. 1.5 with placebo.

(Sandford, 2001) – It is effective in the short-term treatment of panic disorder
  • 16 patients with panic disorder with agoraphobia and at least two full panic attacks in a 2-week period.
    • Average number of full panic attacks was 5.5 per two weeks.
  • Double-blind, randomized, crossover style comparing it to placebo.
    • Pagoclone: 0.1 mg three times daily.
  • Primary outcome measure: frequency of panic attacks measured from the panic attack diary.
  • Results
    • 14/16 completed the study. One was lost to follow-up and one discontinued due to an exacerbation of panic/anxiety during a washout period following treatment with placebo.
    • There was a significant decline in the number of full panic attacks from baseline with pagoclone, but not with placebo. The results were significant if looking at completers or all participants.
      • Significant decreases were not seen in the direct comparison during either trial arm, but the pooled results were significant. Average of 1.5 fewer panic attacks in placebo and 2.2 fewer in pagoclone.
    • No consistent trends on any other outcome measure, including HAM-A, HAM-D, CGI, or Rickels withdrawal scale.
    • No withdrawal symptoms after two weeks of treatment were seen.
    • Treatment with pagoclone did not result in any abnormalities based on ECG, vitals, clinical chemistry, hematology, or urinalysis. No serious adverse effects were reported and the adverse event profiles were similar between pagoclone and placebo.
  • Discussion
    • It’s important to be cautious of the findings because there were no other differences for the secondary measures of anxiety. But, the study suggests that, at least in the short term, pagoclone is well-tolerated and has a much better side effect profile than classical benzodiazepines.
  • COI: Funded by Interneuron Pharmaceuticals.



Some of the interest in the substance has come from its potential use as an anxiolysis-selective GABAergic or as an alcohol alternative in social settings. A few reports of people using 5-30 mg can be found online, but there are too few reports to characterize the effects of those doses. It can only be said that pagoclone appears useful for anxiety and sleep for some people. It doesn’t seem to have much euphoric potential.

Very strong sedative effects don’t appear to exist even at high doses around 5 mg. Diazepam and other benzodiazepines may be more effective in that regard.

An abuse potential study found 4.8 mg of pagoclone significantly increased some measures of drug-liking and “good effects,” while 1.2 mg did not (de Wit, 2006).

(de Wit, 2006) – It has a similar abuse potential to diazepam at higher doses
  • Participants received pagoclone at 1.2 mg (the upper therapeutic dose) and at a supratherapeutic dose of 4.8 mg. Diazepam was given at 30 mg, which is above the medical range for an acute dose.
    • They received the drugs in a double-blind fashion.
  • 25 participants, though only 23 finished because two were non-compliant.
  • Results
    • Measures of drug liking and euphoria
      • Both pagoclone 4.8 mg and diazepam increases ratings of drug liking significantly more than placebo, while pagoclone 1.2 mg did not differ from placebo.
      • Though diazepam significantly increased ratings of willingness to take the drug again, neither pagoclone dose did.
      • Both diazepam and pagoclone 4.8 mg increased ratings of “drug liking” and “good effects,” whereas pagoclone 1.2 mg did not.
      • Drug identification
        • 83% correctly identified diazepam as sedative, 54% identified pagoclone 4.8 mg as sedative, 37% identified pagoclone 1.2 mg as sedative, and 28% identified placebo as sedative.
    • Adverse effects
      • Both diazepam and pagoclone 4.8 mg produced several adverse effects, including ratings of “bad effects.” They also raised the scores on the LSD (dysphoria) scale, which measures dysphoria and somatic effects.
    • Sedation
      • Diazepam and pagoclone (1.2 and 4.8 mg) raised ratings of sedation. But diazepam was significantly more effective than either pagoclone dose.
      • Diazepam also increased ratings on a “sleepy” item relative to placebo and both pagoclone doses. Neither dose of pagoclone significantly increased that measure relative to placebo.
    • Somatic effects
      • Diazepam and pagoclone 4.8 mg caused blurred vision, feelings of being limp/loose, and lightheadedness.
      • Pagoclone 4.8 mg also caused numbness/tingling.
    • Stimulant effects
      • Diazepam and pagoclone 4.8 mg decreased ratings of stimulant effects.
    • Psychomotor impairment
      • Based on the DSST and HVLT Immediate and Delayed Recall tests, diazepam and pagoclone (both doses) produced performance impairment vs. placebo. The impairment was significantly greater from pagoclone 4.8 mg than from 1.2 mg or diazepam.
      • On both the Immediate and Delayed Recall items of the HCLT, subjects recalled fewer words from the memorized list after diazepam or pagoclone (either dose) vs. placebo.
  • Discussion
    • On most of the measures of abuse potential, pagoclone at higher doses had a similar abusability profile to diazepam, while the lower dose in the therapeutic range did not.
    • Both drugs also produced some effects that could limit their abuse potential, such as increased ratings of bad effects and increased scores on the ARCI-LSD scale, which is an indicator of dysphoria and unpleasant somatic effects.
      • Notably, pagoclone seemed to increase LSD scores more than diazepam.
  • COI: The paper was supported by Pfizer.



Pagoclone is a member of the cyclopyrrolone class of nonbenzodiazepines, which also includes the sleep medication zopiclone.

Its structure consists of a core isoindolin-1-one ring-system with a secondary 7-chloro-1,8-naphthyridin-2-yl ring-system bound at its 2 position and a 5-methyl-2-oxohexyl side-chain bound at its 3 position.

It is the active (+)-enantiomer of the racemate RP 59037.


It is a benzodiazepine site agonist. Like the nonbenzodiazepine Z-drugs (e.g. zopiclone and zolpidem), it is more selective than benzodiazepines, meaning it preferentially affects GABAA receptors with specific compositions.

Pagoclone has a high affinity for GABAA receptors with an α1, α2, α3, or α5 subunit (Atack, 2006). It’s essentially a full agonist at α3 receptor subtypes, while it’s a partial agonist at the rest. The EC50 values for those receptor types range from 3.1 to 6.6 nM.

Physical dependence may not develop as quickly, though the evidence for this only comes from a couple animal studies in which the racemate RP 59037 didn’t produce hypersensitivity to a benzodiazepine site inverse agonist after heavy dosing for three days (Stutzmann, 1991; Piot, 1992).

In humans, a PET study supported its status as a partial benzodiazepine site agonist. It reduced the uptake of the benzodiazepine antagonist flumazenil and occupied more receptors than lorazepam while producing less of a response (Lingford-Hughes, 2005).

5′-hydroxy-pagoclone, an active metabolite, may be responsible for more activity than pagoclone itself. It’s present at even higher concentrations in the plasma and brain of rats following pagoclone administration (Atack, 2006) and it’s more potent as a GABAA benzodiazepine site agonist. The metabolite is also a full agonist at receptors with the α1 subunit, unlike pagoclone.

Low doses (0.6 mg/d for five days) combined with ethanol produced no clear synergism (Haig, 2003).

(Atack, 2006) – 5′-hydroxy-pagoclone occurs at higher concentrations in animals and is likely responsible for the effects
  • Research
    • In vitro properties of pagoclone
      • Affinity at GABAA receptors containing different α subunits.
        • High affinity exists for the diazepam-sensitive GABAA complexes while affinity is low for the insensitive receptors containing α4 and α6 subunits.
        • RP 59037
          • α1
            • 1.5 nM
          • α2
            • 4.1 nM
          • α3
            • 1.4 nM
          • α4
            • Over 1000 nM
          • α5
            • 10.5 nM
          • α6
            • Over 1000 nM
        • Pagoclone
          • α1
            • 0.9 nM
          • α2
            • 2.8 nM
          • α3
            • 0.7 nM
          • α4
            • Over 1000 nM
          • α5
            • 9.1 nM
          • α6
            • Over 1000 nM
      • Efficacy
        • Pagoclone has significant agonist activity at GABAA complexes with α1, 2, 3, and 5 subunits. It’s essentially a full agonist at the α3 subtype and is a partial agonist at the others.
        • EC50 values ranged from 3.1 to 6.6 nM.
  • In vivo properties of pagoclone
    • It has anxiolytic-like activity in the rate elevated plus maze
      • It causes a dose-dependent increase in time spent in the open arms but that’s only achieved at the highest dose of 3 mg/kg (oral), which gives a % time in the open arms of 34% vs placebo’s 16%. The comparison drug chlordiazepoxide was significantly more effective.
      • The total distance traveled parameter (an indirect measure of sedation) showed a significant effect. Pagoclone was associated with a significant decrease in distance traveled at all doses tested, indicating it is sedating even at doses (0.3 and 1 mg/kg) that don’t produce notable anxiolytic action.
  • In vivo receptor occupancy (rat brain benzodiazepine binding sites) for pagoclone
    • 0.3 mg/kg: 35%
    • 1 mg/kg: 52%
    • 3 mg/kg: 66%
    • For comparison, the anxiolytic actions of 5 mg/kg (IP) chlordiazepoxide were achieved at a much lower occupancy rate of 23%.
  • Impairs performance in the rat chain-pulling test
    • Response rate data
    • Placebo: 77%
    • Diazepam (10 mg/kg oral): 25%
    • Pagoclone
      • 1 mg/kg: 54%
      • 3 mg/kg: 40%
      • 10 mg/kg: 22%
    • Pagoclone clearly impairs task performance.
  • Pagoclone is metabolized in vivo (in rats) to 5′-hydroxy-pagoclone
    • Drug measurement data revealed the metabolite, in both brain and plasma, exceeds the concentration of pagoclone by 10-20-fold.
    • For example, at 3 mg/kg, the concentrations are:
      • Brain
        • Pagoclone: 2.5 ng/g
        • 5′-hydroxy-pagoclone: 49 ng/g
      • Plasma
        • Pagoclone: 2.2 ng/mL
        • 5′-hydroxy-pagoclone: 26 ng/mL
  • In vitro properties of 5′-hydroxypagoclone
    • It binds to human recombinant GABAA receptors with an affinity of 0.6-3.6 nM, marginally higher than pagoclone itself. It’s also a full agonist at the α3 site (147%) and a partial agonist at α2 (46%) and α5 subtypes (60%).
      • In contrast to pagoclone, which was a partial agonist at the α1 subtype (66%), the metabolite was a full agonist with a 150% effect.
  • In vivo properties of 5′-hydroxy-pagoclone
    • It has anxiolytic-like effects in the rat elevated plus maze
      • All three doses (0.3, 1, and 3 mg/kg IP) significantly increased open arm time relative to placebo.
      • Like pagoclone, it decreased the total distance traveled, in contrast to chlordiazepoxide.
    • Occupancy
      • 0.3 mg/kg: 40%
      • 1 mg/kg: 67%
      • 3 mg/kg: 87%
    • It reduces locomotor activity
      • It significantly dose-dependently decreased activity.
  • Discussion
    • Though pagoclone has attracted attention due to its purported anxioselective properties, it may not be selective. Though a Phase 1 study in humans did indicate a separation between sedative and anxiolytic effects.
    • Little metabolic data is available from human subjects. But data from squirrel monkeys shows 5′-hydroxy-pagoclone is present at higher concentrations than pagoclone and a PET study in humans showed, though without qualitative elaboration, that it is present in humans.
  • COI: The paper came from Merck’s Neuroscience Research Center.
(Lingford-Hughes, 2005) – PET data indicates pagoclone is a partial agonist
  • Labeled flumazenil PET was used to measure benzodiazepine site occupancy. At the same time, saccadic eye movements were measured since benzodiazepines have a reliable effect on eye movements.
  • Lorazepam (1 mg oral) was used as the full agonist, while pagoclone was given at 0.4 mg.
  • 6 healthy male volunteers were used. The drugs were given about 90 min before the injection of flumazenil.
  • Results
    • Both drugs reduced flumazenil uptake, consistent with them occupying benzodiazepine receptors. Pagoclone occupied more than lorazepam in the frontal cortex (mean of 14.7% vs 5.6%), in the thalamus (mean 13.0% vs. 8.9%), and in the cerebellum (16.6% & 6.17% for the pagoclone participants vs. 3.9%, 4.7%, and 13.9% for the lorazepam group).
    • There were no significant differences in subjective anxiety or sedation between the drugs.
    • Lorazepam impaired all of the saccadic eye movement parameters calculated. While pagoclone also impaired the parameters, it was less effective, such as producing either less deceleration or acceleration of eye movements while lorazepam consistently caused deceleration.
    • When comparing the pharmacodynamic effects to receptor occupancy in the frontal cortex, lorazepam showed a much greater effect.
  • Discussion
    • This evidence is consistent with pagoclone being a partial agonist. The low % occupancy by lorazepam is consistent with previous studies estimating 95% of benzodiazepine receptors can be unoccupied by a full agonist while still having full behavioral potency.
    • For comparison, significant drowsiness and sleep with benzodiazepines is associated with 30% occupancy, and sleepiness from zolpidem 20 mg is associated with 21% occupancy.
  • COI: The study was funded by Interneuron.
(Haig, 2003) – Pagoclone and ethanol have no synergism for impairment and aren’t highly additive
  • 16 healthy females.
  • Pagoclone was given at 0.6 mg twice daily for 5 days. After the last pagoclone dose, ethanol 0.7 g/kg or placebo was administered.
  • Results
    • The incidence of CNS and GI adverse events was high and similar between ethanol and ethanol/pagoclone.
    • The treatment rank for neuropsychometric performance (worst to best) was: pagoclone/ethanol > ethanol > pagoclone > placebo.
    • Pagoclone did not have synergistic effects with ethanol. It did have additive effects in motor screening and in part of a memory test.
    • There was no pharmacokinetic interaction.
  • Discussion
    • This study shows no pharmacokinetic or synergistic pharmacodynamic interaction exists between the drugs and pagoclone exhibited few additive effects.
  • COI: Not reported
(Wong, 1995) – The racemate RP 59037 is a partial agonist at diazepam-sensitive sites
  • Effect of ligands on diazepam-insensitive and diazepam-sensitive Ro 15-4513 binding sites in rat cerebellar and cortical membranes
    • Affinity (Ki) for diazepam-insensitive sites
      • Over 3000 nM
    • Affinity (Ki) for diazepam-sensitive sites
      • 0.3 nM
    • Maximal displacement
      • Alcohol-sensitive rats (have a single point mutation in the GABAA receptor α6 subunit that makes diazepam-insensitive sites become sensitive to benzodiazepine agonists)
        • 98%
      • Alcohol-insensitive rats
        • 82%
  • COI: Not reported
(Jackson, 1992) – Suriclone behaves like a full agonist, while RP 59037 behaves like a partial agonist based on their respective impacts on mice body temperature.
  • Mice
    • RP 59037 was given at 3, 10, and 30 mg/kg IP
  • Results
    • The full agonist suriclone (3, 10, and 30 mg/kg) produced significant hypothermia in a dose-dependent manner. The maximal fall in body temperature from 10 and 30 mg/kg was around 3-4°C. Flumazenil (10 mg/kg), a benzodiazepine antagonist, blocked the decrease in temperature.
    • RP 59037 had no effect on temperature at 3 mg/kg. At higher doses it produced a significantly lower temperature, but only with maximal responses of 1-1.5°C.
      • At 10 mg/kg it did not significantly antagonize the hypothermic effect of suriclone 30 mg/kg.
  • Discussion
    • RP 59037 is much less effective at producing hypothermia in mice than suriclone, which fits with its partial agonist profile. Its maximal impact on body temperature was only around one-third that of suriclone.
    • However, RP 59037 also differs from other partial agonists that have been shown to have no hypothermic property in this strain of mice. So it could be more efficacious than other partial agonists in this model, and that could explain why it did not reduce the hypothermic effect of the full agonist suriclone.
  • COI: Not reported
(Piot, 1992) [overlapping data with (Stutzmann, 1991)] – The racemate RP 59037 is a partial GABAA agonist and is more selective than diazepam with its observed effects
  • In vitro
    • It displaced the binding of flunitrazepam to rat cortical membranes with an IC50 of 1.6 nM. It enhanced the binding of TBPS, a GABAA antagonist, to the picrotoxin site by 37%.
  • In vivo
    • It displaced the binding of flumazenil in the rat cerebral cortex with an ID50 of 3.51 mg/kg (oral).
    • It markedly reduced anxiety in three rat models.
      • In the plus-maze test, it significantly increased anxiolytic-like activity starting from 0.63 mg/kg (oral).
    • It antagonized the discriminative stimulus provided by pentylenetetrazole at an ED50 value six times lower than diazepam (0.28 vs. 1.77 mg/kg oral).
    • It strongly antagonized pentylenetetrazole-induced convulsions (ED50 = 0.21 mg/kg in mice and 0.59 mg/kg in rats). Though it did not protect against picrotoxin-, strychnine-, or supramaximal electroshock-induced convulsions in mice.
    • The drug was practically devoid of CNS depression. It impaired performance much less than diazepam in the rota-rod and inclined screen tests in rats.
    • At 800 mg/kg (oral) it neither interacted with hexobarbital nor potentiated the effects of ethanol, unlike diazepam.
    • Withdrawal
      • Rats withdrawn from any dose from 2 to 400 mg/kg (IP) daily did not see potentiation of FG7142-induced convulsions, whereas potentiation was seen with diazepam withdrawal.
  • COI: Authors are from the Rhone-Poulenc Rorer company.

(Stutzmann, 1991) – RP 59037, the racemate, is a partial GABAA agonist.

  • In mice, it (like diazepam) protects against convulsions induced by agents active at the GABA receptor-complex, such as pentylenetetrazole (ED50: 0.21 mg/kg oral), isoniazid (0.26 mg/kg), bicuculline (0.087 mg/kg), and 3-mercaptopropionic acid (2.6 mg/kg).
  • In an operant conflict procedure in rats it was more potent than the reference drug, alprazolam. Their minimal effective doses for reversing conflict-induced inhibition of drinking behavior were 0.325 mg/kg and 1.25 mg/kg oral, respectively.
  • Withdrawal from subchronic treatment (doses of up to 100 mg/kg IP four times daily for 3 days) did not result in hypersensitivity to FG7142.
  • COI: Authors are from the Rhone-Poulenc Rorer company.


Human and animal studies have shown pagoclone is extensively metabolized to the active substance 5′-hydroxy-pagoclone (Dalvie, 2009; Atack, 2006).

(Dalvie, 2009) – Pagoclone is metabolized to 5′-hydroxy-pagoclone to a significant extent in humans

  • Data from Pfizer’s database.
  • 5′-hydroxy-pagoclone is a phase 1 metabolite and accounts for 65% of circulating radioactivity in humans.
  • Incubation studies showed 5′-hydroxy-pagoclone could be detected in hepatocytes and S-9 factions (both 100% successfully), but not in liver microsomes.
  • COI: The paper came from Pfizer.



Pagoclone was discovered by Rhone-Poulenc, a French pharmaceutical company that would come to be known as Aventis following a merger with Hoechst Marion Roussel. A 1994 patent (5498716) from Rhone-Poulenc described it as exhibiting “remarkably anxiolytic, hypnotic, anticonvulsant, antiepileptic, and muscle relaxant properties.”

Rhone-Poulenc licensed it to Interneuron Pharmaceuticals, an American pharmaceutical company that later changed its name to Indevus.

December 1999

Interneuron (Indevus) licensed the substance to Warner-Lambert, which was acquired by Pfizer a few months later. Pfizer received exclusive worldwide rights to pagoclone for the treatment of panic and generalized anxiety disorders. In exchange, Interneuron received an upfront payment of $13.75 million and was entitled to receive up to $60 million in additional payments based on clinical and regulatory milestones, as well as royalties from net sales.

Under this deal, Pfizer would conduct and fund all further clinical development, regulatory review, manufacturing, and marketing for the substance.

Pagoclone is one of Interneuron’s most valuable assets, and we are enthusiastic that our partner, Pfizer, has initiated the clinical program on this important product…We continue to be very impressed with Pfizer ‘s extensive capabilities in taking pagoclone forward in a comprehensive and timely manner, reflecting the recognition by both companies that pagoclone has the potential to make a significant impact upon the market to treat panic disorder and other anxiety disorders. – Dr. Glenn Cooper, president and CEO, Interneuron Pharmaceuticals



Indevus received the exclusive worldwide rights to pagoclone back from Pfizer, which decided not to pursue development of the drug for anxiety and panic conditions. This came after insufficient efficacy was demonstrated in those conditions. Because Aventis had some rights to the drug, Indevus and Aventis would decide together about pagoclone’s future following additional analysis of trial data.

While we are disappointed that Pfizer has elected not to pursue this project, we believe that further development of pagoclone for anxiety disorders may be merited, – Dr. Glenn Cooper, Indevus


Indevus announced it had begun new corporate partnership discussions and it would decide on the clinical development of the drug for anxiety/panic based on analyses of six clinical trials and based on ongoing consultation with Aventis.


Due to its contractual rights to the drug, Aventis had a period of 90 days from the termination of the Pfizer deal to decide if it would develop the drug. It decided not to.

Early 2005

Indevus received a US patent covering the use of pagoclone for the treatment of stuttering.



An article in the Boston Globe said pagoclone “shows promise” for stutterers. Brooke Wagner, an Indevus spokesperson, said a larger trial with 600 to 1,000 patients was needed to confirm its effects.


ABC News reported that pagoclone, unlike other stuttering medications, produces just a few mild side effects and researchers believe it works by lowering dopamine levels in the brain.

Some perspectives on the use of pagoclone or other medications for stuttering were featured:

Medical therapy is not likely to be a first-line treatment for stuttering. – Richard Merson, a speech and language pathologist at William Beaumont Hospital in Royal Oak, Michigan

Merson said it will likely only be used in the tiny minority of patients who are not helped by speech therapy.

As a stutterer, I see the value in medication but also the importance of good speech therapy, – Ross Barrett, a speech therapist at East Virginia School of Medicine


Indevus announced it was moving towards regulatory approval for persistent developmental stuttering (PDS) and that it would initiate a Phase 3 trial in the first half of 2007. The company intended to explore pediatric uses as well and it would start a small pharmacokinetic population in children so that it could pick doses for a Phase 2 or 3 study in 2007.

The Company met with the FDA to discuss the results of our Phase II trial in adults with PDS and to explore ideas for the design and conduct of future trials…Although the FDA had never considered a drug for stuttering, FDA officials were, in my opinion, extremely well-prepared and were able to give us specific and useful guidance that has allowed us to map out a clear path toward an NDA submission.

Specifically, the FDA advised the Company to: 1) pursue pediatric studies in parallel with adult studies so that if pagoclone is effective and safe in both populations, the NDA could be approvable for the broadest possible stuttering population; 2) conduct the next adult and pediatric placebo-controlled trials as fixed dose-response studies to determine the minimally-effective dose; 3) pursue Phase III trials under special protocol assessments (SPAs) to allow the FDA to formally sign off on trial designs.

Importantly, the FDA tentatively agreed, pending the SPA review process, to the Company’s proposal on primary and secondary study endpoints and study duration. – Dr. Glenn Cooper, CEO, Indevus

It also announced it would not continue a Phase 2 trial for premature ejaculation due to insufficient efficacy.

September 2008

Interest in the drug grew after a promising Phase 2 trial. Indevus signed a development, licensing, and commercialization agreement with Teva Pharmaceuticals, which received the exclusive, worldwide rights to pagoclone.

Indevus would conduct the Phase 2B study and be reimbursed by Teva. The new placebo-controlled trial would involve 300 patients and a six-month evaluation period. Enrollment was set to begin by Q1 2009.



Though it was the first drug tested through the FDA process specifically for stuttering, Dr. Gerald Maguire, one of the main pagoclone researchers, revealed the latest data was not impressive. During the US National Stuttering Association conference in Ohio, Dr. Maguire informally discussed the Phase 2B study results and said the pre-specified criteria for success were not met. Significant and lasting gains in fluency were not seen.


Dr. Gerald Maguire said pagoclone would not be approved by the FDA for stuttering anytime soon. He thinks the antipsychotic asenapine “is going to be better.”


Endo Pharmaceuticals decided not to continue its research program focused on pagoclone for stuttering.


Some anecdotal reports of success with pagoclone for stuttering were seen in the stuttering community online and it was also sold via research chemical websites, though it’s rarely been used outside of trial settings.

As of 2018

It has never been commercialized and is an uncommon drug.

Legality (as of June 2018)

United States: Unscheduled

Australia: Not specifically controlled

Canada: Unscheduled

UK: It is not specifically scheduled, but it does fall under the Psychoactive Substances Act.


At the low medical doses of 0.3 to 0.6 mg/d, the side effects are minimal and not very concerning. It can cause drowsiness, headache, and impairment in some individuals, but that’s about the extent of the concern based on the available data.

Even at higher doses, such as the 4.8 mg used in the abuse potential study, the primary issue is impairment. A direct toxic effect capable of causing death is unlikely to be reached at any of the doses described in this overview. There may be a lethal dose, but pagoclone does not have a low therapeutic index and seems to be a mostly nontoxic drug at the doses evaluated so far.

Physical dependence may build slower than with benzodiazepines, but it probably does build over a sufficient length of time, leading to tolerance and perhaps withdrawal symptoms like anxiety and insomnia.


Pagoclone can acutely cause sedation and impairment in learning, memory, and alertness.

(Caveney, 2008) – It impairs some aspects of performance, though tolerance removes the effect.
  • 12 healthy subjects. Open label, randomized, multiple-dose, crossover study
  • They were given a low 0.15 mg dose, a medium 0.30 mg dose, and a high 0.60 mg dose every 12 hours for 7 days followed by a 7-day washout between periods.
  • Results
    • Sleepiness and alertness
      • At all doses subjects had a significantly greater increase in fatigue on Day 1 vs. baseline than from Day 6 vs. baseline. The medium dose was associated with significantly more fatigue than the low dose.
    • Learning and memory
      • There was a consistent pattern of effects on the Buschke Selective Reminding Test (i.e. Immediate Recall, LTS, and Delayed Recall) being localized to Day 1, with Day 6 scores essentially returning to baseline.
      • The decline from baseline in Immediate Recall was significantly greater at the high vs. low dose.
      • For the LTS (Long Term Storage) score, significantly greater declines from baseline were seen at both the medium and high doses vs. the low dose.
      • For Delayed Recall, no change was seen between the low and medium dose, but the decline at the high dose was significantly greater than the other two doses.
      • Overall, there was a pattern of no change or improvement at the low dose, with declines in performance for the other doses.
    • Simple/Complex attention
      • There was no effect of Day or Dose on Reaction Time or Movement Time for the Simple Reaction Time measures.
      • For the Choice Reaction Time measures, there was no effect of Day or Dose on reaction time, but there was a main effect of Dose on Movement Time.
    • Psychomotor problem solving
      • Subjects had significantly improved test scores on Day 6 vs. Day 1. There was essentially no change in performance on Day 1 vs. baseline, but on Day 6 there was a rise in performance compared to baseline, though the mean change was not very large.
  • Discussion
    • Decrements in performance for leaning and memory were generally greater for the high dose compared to the low or medium dose.
    • The speed and fatigue scores seemed to improve at the low dose, while declining at the medium dose and declining to a lesser extent at the high dose. But the clinical significance of this is not clear. Across doses, the largest mean change did not exceed one point on the sleepiness scale, which means the change would not usually be viewed as clinically significant.
    • The sedative effects appear to be only very mild in nature, especially since objective measures of motor speed and attention were not evident.
    • Learning and memory scores did suggest a consistent pattern of lower performance at the high dose, but by Day 6 the changes were no longer evident.
  • COI: Supported by Pfizer

Risky Combinations (list may not be exhaustive)

Other depressants including alcohol (ethanol), benzodiazepines, and opioids.


Atack, J. R., Pike, A., Marshall, G., Stanley, J., Lincoln, R., Cook, S. M., … Reynolds, D. S. (2006). The in vivo properties of pagoclone in rat are most likely mediated by 5′-hydroxy pagoclone. Neuropharmacology, 50(6), 677–689.

Caveney, A. F., Giordani, B., & Haig, G. M. (2008). Preliminary effects of pagoclone, a partial GABA(A) agonist, on neuropsychological performance. Neuropsychiatric Disease and Treatment, 4(1), 277–82. Retrieved from

Dalvie, D., Obach, R. S., Kang, P., Prakash, C., Loi, C., Hurst, S., … Smith, D. A. (2009). Assessment of Three Human in Vitro Systems in the Generation of Major Human Excretory and Circulating Metabolites. Chemical Research in Toxicology, 22(2), 357–368.

de Wit, H., Vicini, L., Haig, G. M., Hunt, T., & Feltner, D. (2006). Evaluation of the Abuse Potential of Pagoclone, a Partial GABAA Agonist. Journal of Clinical Psychopharmacology, 26(3), 268–273.

Haig, G. M., Giordani, B., Randinitis, E. J., & Mitchell, D. Y. (2003). Evaluation of the pharmacodynamic and pharmacokinetic interaction between pagoclone and ethanol. Clinical Pharmacology & Therapeutics, 73(2), P93–P93.

Ingham, R. J. (2010). Comments on Article by Maguire et al: Pagoclone Trial. Journal of Clinical Psychopharmacology, 30(5), 649–650.

Jackson, H. C., Ramsay, E., & Nutt, D. J. (1992). Effect of the cyclopyrrolones suriclone and RP 59037 on body temperature in mice. European Journal of Pharmacology, 216(1), 23–7. Retrieved from

Lingford-Hughes, A., Wilson, S. J., Feeney, A., Grasby, P. G., & Nutt, D. J. (2005). A proof-of-concept study using [11C]flumazenil PET to demonstrate that pagoclone is a partial agonist. Psychopharmacology, 180(4), 1–3.

Maguire, G. A., Yeh, C. Y., & Ito, B. S. (2012). Overview of the Diagnosis and Treatment of Stuttering. Journal of Experimental & Clinical Medicine, 4(2), 92–97.

Maguire, G., Franklin, D., Vatakis, N. G., Morgenshtern, E., Denko, T., Yaruss, J. S., … Riley, G. (2010). Exploratory Randomized Clinical Study of Pagoclone in Persistent Developmental Stuttering. Journal of Clinical Psychopharmacology, 30(1), 48–56.

Reibaud, M. (1992). Rp 59037 : a Cyclopyrrolone With Partial Agonist Properties At the Gaba Receptor Complex. Journal of Psychopharmacology, 6(1), 124–124.

Sandford, J. J., Forshall, S., Bell, C., Argyropoulos, S., Rich, A., D’Orlando, K. J., … Nutt, D. J. (2001). Crossover trial of pagoclone and placebo in patients with DSM-IV panic disorder. Journal of Psychopharmacology, 15(3), 205–208.

Stutzmann, J. ., Piot, O., Rataud, J., Bardone, M. ., & Blanchard, J. . (1991). Behavioural and electrocorticographic studies in rodents with RP 59037, a cyclopyrrolone derivative partial agonist at the GABA receptor-complex. European Neuropsychopharmacology, 1(3), 397.

Wong, G., Uusi-Oukari, M., Hansen, H. C., Suzdak, P. D., & Korpi, E. R. (1995). Characterization of novel ligands for wild-type and natural mutant diazepam-insensitive benzodiazepine receptors. European Journal of Pharmacology, 289(2), 335–42. Retrieved from

Category Tag