Published on 15 September 2026
Mirtazapine and methamphetamine: what does the first phase 3 trial of this drug actually show?
In brief
At the date this trial was published, in 2026, no medicine holds a marketing authorisation for methamphetamine use disorder. The Tina trial, conducted between 16 November 2022 and 1 May 2025 in six Australian outpatient alcohol and other drug clinics, is the first phase 3 trial of mirtazapine in this indication. Of 344 participants randomised, 339 received the intervention, 172 on mirtazapine 30 mg daily for twelve weeks and 167 on matched placebo. Mean age was 42.0 years (standard deviation 8.6), 126 participants (37.2%) were female, and median use at baseline reached 24 of the previous 28 days (interquartile range 17 to 28). The primary end point, the change in self-reported days of use over 28 days between baseline and week 12, is met: the mean reduction is 7.0 days on mirtazapine against 4.8 days on placebo, a mean difference of 2.2 days (95% CI −4.2 to −0.2; P = .02). No secondary end point is significant, including the only biological measure in the trial, the proportion of methamphetamine-negative oral fluid samples. On tolerability, drowsiness affects 47% of patients on mirtazapine against 33% on placebo, weight gain 10% against 3%, and 40 participants (23%) discontinued mirtazapine for an adverse event against 25 (15%) on placebo. The authors conclude that mirtazapine, delivered in routine clinical practice, reduces methamphetamine use, with no unexpected safety signal.
The context
Three things have to be set down before reading the figures, because they determine what this trial really demonstrates.
First point: this is neither the first positive trial, nor the first phase 3 trial conducted in this disorder. It is the first phase 3 trial of mirtazapine. The distinction is not cosmetic. Two phase 2 randomised controlled trials had already reported a reduction in use on mirtazapine 30 mg daily: Colfax and colleagues in 2011 (60 participants, relative risk of a positive urine test 0.57; 95% CI 0.35 to 0.93; P = .02) and Coffin and colleagues in 2020 (120 participants, relative risk 0.67 at week 12; 95% CI 0.51 to 0.87). And another phase 3 trial preceded Tina in this disorder: ADAPT-2 (NIDA CTN-0068 protocol, NCT03078075), which enrolled 403 participants and tested extended-release injectable naltrexone combined with bupropion, with a biological primary end point, at least three of four negative urine samples at the end of the stage, and which was published as positive in the New England Journal of Medicine in 2021: weighted response 13.6% on active treatment against 2.5% on placebo. The priority claim made in the Tina protocol is narrower than that, and it has to be quoted with its restriction: Tina is “the first phase 3 placebo-controlled randomised trial to examine the effectiveness and safety of mirtazapine as an outpatient pharmacotherapy for methamphetamine use disorder”. Writing that no medicine had ever shown efficacy in this disorder would therefore be inaccurate. What is accurate, and what the published abstract states, is that no medicine holds an approval for it.
Second point: the two earlier mirtazapine trials did not study the same population, and neither one is an ordinary clinical caseload. Colfax 2011 recruited men who have sex with men; Coffin 2020 recruited cisgender men and transgender women who have sex with men. Both were recruited in San Francisco, with HIV risk reduction as the leading concern. The dosing schedules also differed: Colfax 2011 gave 30 mg from the outset, without titration, whereas Coffin 2020 allowed one week at 15 mg before moving to 30 mg through to week 24. Tina uses a schedule without titration, which bears directly on how its tolerability should be read. A 2022 meta-analysis pooled these two trials: a small and probable reduction in use at twelve weeks, with a relative risk of 0.81 whose 95% confidence interval (0.63 to 1.03) crossed the null, in 133 participants, at moderate certainty. That 0.81 does not compare term for term with the 0.57 and 0.67 of the individual trials: the meta-analysis pools a risk of use at twelve weeks covering 133 of the 180 participants in the two trials, whereas each trial reported the risk of a positive urine test under its own analytical convention. The gap between these three values is therefore not an arithmetical inconsistency. What Tina adds is not the direction of the effect, already suggested, but its assessment on a larger scale, in an ordinary outpatient population of which 37% are women, and under routine practice conditions.
Third point: the epidemiological question does not arise in the same terms everywhere. Treatment demand for methamphetamine is very unevenly distributed across countries, and most of the figures that circulate on this subject cannot be verified. The question is handled separately, further down.
The study at a glance
| Question (PICO) | |
|---|---|
| Population | |
| Adults with moderate to severe methamphetamine use disorder, recruited in 6 Australian outpatient alcohol and other drug treatment services. 344 randomised, 339 treated. Mean age 42.0 years (standard deviation 8.6), 37.2% female, median use at baseline of 24 of 28 days (interquartile range 17 to 28). The eligibility age range, 18 to 65 years, comes from the published protocol and not from the abstract of the article | |
| Intervention | |
| Mirtazapine 30 mg daily for 12 weeks, as a take-home medication under the protocol. The published protocol sets out the schedule: one 30 mg tablet in the evening from the outset, with no titration phase, with a possible reduction to 15 mg, temporary or permanent, in the event of an adverse reaction, then a taper to 15 mg daily for 28 days from week 12 | |
| Comparator | |
| Matched placebo | |
| Primary end point | |
| Change in the number of days of methamphetamine use over the previous 28 days, between baseline and week 12. A self-reported measure, per the published protocol | |
| Secondary end points | |
| Depression, insomnia, HIV risk behaviour, quality of life, methamphetamine-negative oral fluid samples | |
| Design | |
| Phase 3 randomised controlled trial, parallel groups, double-blind, placebo-controlled, multicentre. Enrolment from 16 November 2022 to 1 May 2025, data analysis from May to September 2025. Registered as ACTRN12622000235707 on 9 February 2022 · CEBM 1b |
Quality control
| Item | Status |
|---|---|
| Prospective registration | Sound |
| FindingAustralian and New Zealand registry, ACTRN12622000235707, registered on 9 February 2022, more than nine months before the first enrolment. Full protocol published in 2024, before the end of enrolment | |
| Randomisation and blinding | Sound |
| FindingParallel groups, 1:1 allocation per the protocol, double-blind, matched placebo. Randomisation was stratified by site, sex and PHQ-9 score (below 10 against 10 or more): comparability of the groups on depressive severity at baseline was therefore guaranteed by design, which makes the negative result on the depression secondary end point harder to attribute to an initial imbalance. The design is what a phase 3 trial should look like | |
| Integrity of the blind | Reservation |
| FindingDrowsiness affects 47% of patients on mirtazapine against 33% on placebo, weight gain 10% against 3%. A patient who sleeps and gains weight may guess their arm. No test of the integrity of the blind appears in the published abstract, which does not mean that none was performed | |
| Nature of the primary end point | Reservation |
| FindingThe published protocol defines the primary end point as a number of self-reported days of use. A self-reported measure combined with a potentially imperfect blind is the combination most exposed to differential measurement bias | |
| Biological measure | Weak point |
| FindingThe only objective marker in the trial, a negative oral fluid sample, is classified as a secondary end point and is not significant. The self-reported result is therefore not corroborated by the biological measure | |
| Treatment adherence | Reservation |
| FindingThe protocol provides for adherence monitoring with electronic MEMS caps. No adherence figure appears in the published abstract. Precise values circulate with no identifiable source, they do not appear in the publication and are not repeated here | |
| Retention and missing data | Reservation |
| FindingThe analysis plan is set out in the published protocol: a mixed model with a group by time interaction on the baseline, week 4, week 8 and week 12 measures, a primary intention-to-treat analysis on non-imputed data, sensitivity analyses with imputation of missing data, and a per-protocol analysis excluding non-adherent participants. What is missing is therefore not the planned model, but confirmation that it was applied as written and the follow-up rate actually achieved at week 12. In a twelve-week trial among active users, that last figure carries real weight | |
| Independence | Reservation |
| FindingThe published protocol declares funding from the Australian Medical Research Future Fund, grant no. 2007155, stating that the funder played no part in the design of the study, the collection and analysis of the data, the interpretation of the results or the writing of the manuscripts. That same protocol also details pharmaceutical industry ties for several of the authors. The drug is a generic in the public domain, which removes the promotional interest of a manufacturer, without removing other forms of interest. The interest declarations published with the article itself were not consulted | |
| Stated aim of the sponsors | Reservation |
| FindingThe published protocol announces that a favourable result will support an application for methamphetamine use disorder to be included as a therapeutic indication for mirtazapine. The aim is legitimate and declared. It justifies reading the chosen end point all the more attentively | |
The findings
| End point | Published result |
|---|---|
| Self-reported days of use out of 28, change from baseline to week 12 (primary end point) | Mean reduction of 7.0 days on mirtazapine against 4.8 days on placebo. Mean difference 2.2 days (95% CI −4.2 to −0.2; P = .02) |
| PEB readingPositive the interval excludes zero, but the bound nearest the null sits at 0.2 days per month, that is one fifth of a day of use avoided under the least favourable hypothesis compatible with the data | |
| Methamphetamine-negative oral fluid | Not significant |
| PEB readingNot concordant the only objective check on the self-reported reduction does not confirm it | |
| Depression | Not significant |
| PEB readingNegative the drug is an antidepressant: this result deserves to be remembered by any prescriber tempted to justify the prescription by a mood comorbidity | |
| Insomnia | Not significant |
| PEB readingNegative whereas the H1 antagonism of the drug and the observed rate of drowsiness could have led one to expect it | |
| HIV risk behaviour | Not significant |
| PEB readingNegative yet this was the most striking signal in the two phase 2 trials, conducted in a population at far higher risk | |
| Quality of life | Not significant |
| PEB readingNegative no benefit the patient could perceive is demonstrated | |
| Drowsiness | 47% on mirtazapine against 33% on placebo |
| PEB readingExpected a gap of 14 points, consistent with the pharmacological profile of the drug | |
| Weight gain | 10% on mirtazapine against 3% on placebo |
| PEB readingExpected a gap of 7 points, over twelve weeks only. In patients whose metabolic health is often already impaired, this point calls for monitoring | |
| Discontinuation for an adverse event | 40 participants (23%) on mirtazapine against 25 (15%) on placebo |
| PEB readingReservation close to one patient in four stops treatment. The placebo comparator, at 15%, shows that part of these discontinuations is not attributable to the drug | |
Three points of reading are needed on these figures, because they govern how the trial will be quoted from here on.
On the sign convention. The published abstract writes the mean difference as positive, 2.2 days, and its confidence interval as negative, from −4.2 to −0.2. Point estimate and interval therefore carry opposite signs in the published text. The reading is not in doubt given the reductions reported in each arm, 7.0 against 4.8 days: mirtazapine removes 2.2 days of use more than placebo does. A slightly different interval, from −4.1 to −0.3, circulates with no identifiable source: it does not match the published text.
On what the abstract does not contain. No incidence rate ratio, no percentage reduction, no abstinence rate, no number of days reached at week 12 in each group. Precise values circulate with no identifiable source, 0.71 for the rate ratio and a 29% relative reduction: they do not appear in the published abstract and PEB does not advance them. It should also be noted that it would be a mistake to combine the median use at baseline, 24 days, with the mean reductions of 7.0 and 4.8 days in order to reconstruct a level at twelve weeks. A median and a mean do not add up.
On what 30 mg daily for twelve weeks actually covers. The published protocol provides for one 30 mg tablet in the evening from the start, with no titration phase, allows a reduction to 15 mg, temporary or permanent, in the event of an adverse reaction, and provides for a taper to 15 mg daily for 28 days from week 12. Two consequences follow. The absence of titration sheds direct light on the 47% rate of drowsiness and the 23% of discontinuations for an adverse event: the full dose is reached at once, at the moment when tolerability is at its worst, which sets Tina apart from the Coffin trial, where one week at 15 mg preceded the move to 30 mg. And exposure was not uniform: not every participant in the active arm received 30 mg for twelve weeks, and the published abstract gives no way of knowing how many were brought back to 15 mg.
Critical appraisal
| Domain | Judgement |
|---|---|
| Level of evidence of the design | Sound |
| FindingPhase 3, randomised, double-blind, placebo-controlled, six sites, prospectively registered, protocol published, sample size in line with the stated target of 340. This is the best available level of evidence for mirtazapine in this indication. Against ADAPT-2, the other phase 3 trial conducted in this disorder, Tina falls behind on one precise point: ADAPT-2 rested on a biological primary end point, urine samples, whereas Tina’s is self-reported | |
| Self-report and biology disagree | Weak point |
| FindingThis is the central flaw in the file. The primary end point, self-reported, is positive. The biological marker, secondary, is not. Two explanations coexist and the data do not settle between them: either oral fluid lacks the sensitivity to capture a reduction of two days per month, or the reported reduction is partly an artefact of an imperfect blind. The first hypothesis is plausible, the second cannot be ruled out | |
| Functional unblinding | Reservation |
| FindingDrowsiness and weight gain increased by 14 and 7 points. A patient who guesses that they are receiving active treatment, and who is asked to count their days of use, is exposed to a directional reporting bias. The direction of that bias, if it exists, inflates the observed effect | |
| Magnitude of the effect | Reservation |
| FindingThis is the strongest argument against an enthusiastic reading, and it comes from the trial itself. The published protocol sized the sample, 340 participants, to detect a minimum rate ratio of 0.75, that is a move from 25 to 20 days of use out of 28, roughly five days, with 90% power. The observed effect, 2.2 days, is clearly below the target the trial had set for itself. The bound of the interval nearest the null, 0.2 days, is moreover compatible with a negligible effect. And no validated threshold of clinical relevance exists for this end point: nobody can therefore state, with data in hand, that a reduction of two days per month changes a patient’s trajectory | |
| No benefit on experienced outcomes | Reservation |
| FindingNeither quality of life, nor depression, nor insomnia moves. The improvement is measured on consumption, not on what the patient feels. Twelve weeks may be too short for the second to follow the first, but that remains a hypothesis | |
| Tolerability and feasibility | Reservation |
| Finding23% of discontinuations for an adverse event against 15% on placebo. The drug is old, its profile is known and predictable, which is a real asset in practice. But one patient in four does not reach the end, and the chosen schedule, 30 mg from the outset without titration, is probably not unrelated to that | |
| External validity | Reservation |
| FindingAn Australian population, using on a median of 24 of 28 days, managed in specialist addiction services. Transposing this to an outpatient psychiatric caseload elsewhere is far from straightforward, whether on the pattern of use or on the setting of care | |
| What the abstract does not say | Reservation |
| FindingThe follow-up rate at week 12, confirmation that the analysis plan set out in the protocol was indeed applied, the results of the sensitivity analyses and of the per-protocol analysis, adherence measured by MEMS, the proportion of participants reduced to 15 mg, subgroup results, serious adverse events. These items probably exist in the full text, which was not consulted. Their absence from the abstract is not an absence of data, and the analysis presented here therefore stops at what the published abstract allows one to establish | |
Things have to be named precisely. What is demonstrated: in a properly conducted phase 3 trial, mirtazapine 30 mg daily for twelve weeks reduces the number of self-reported days of use more than placebo does, by 2.2 days out of 28, with tolerability in line with its known profile and a discontinuation rate higher than placebo. What is suggested: that this reduction corresponds to a real change in behaviour, a hypothesis supported by the agreement in direction with the two phase 2 trials and by pharmacological plausibility, but which the biological measure in this trial does not confirm. What belongs to expert opinion: that a reduction of two days per month justifies exposing one patient in four to a discontinuation for an adverse event, and that treatment continued beyond twelve weeks would eventually produce a benefit on quality of life. No data from this trial settle these last two questions.
A word, finally, on the argument heard everywhere, the old medicine rediscovered. It is appealing and it is partly false. Mirtazapine was not rediscovered in 2026: it has been tested in this indication since at least 2007, the year the first randomised trial, published in 2011, began. It is often argued, to explain those fifteen years, that what was missing was not the idea but the funding for a trial of sufficient size on a drug whose patent nobody holds any more. That is a hypothesis, and ADAPT-2 weakens it: that publicly funded trial of 403 participants also studied generic drugs, and it was carried through. The observable fact is that Tina was funded by an Australian public fund and not by a manufacturer.
Level of evidence
PEB appraisal: high confidence in the experimental design, moderate confidence in the primary result, low confidence in what it means clinically for the patient. The design is that of a serious, prospectively registered phase 3 trial, which places this work well above everything that existed on mirtazapine in this indication. Confidence in the primary result is lowered by one precise point and not by a reservation of principle: the end point is self-reported, the blind is weakened by recognisable adverse effects, and the only biological marker in the trial does not corroborate the result. Confidence in the clinical meaning is lower still, because no outcome perceived by the patient moves and because no threshold of relevance exists against which to judge 2.2 days out of 28.
The colleague test
What an experienced colleague would say if you put this study to them in two minutes, between two consultations.
“ First phase 3 of mirtazapine in this disorder, and there is still no approved medicine: I will take it. But the effect is two fewer days of use per month, when the trial was calibrated to detect five, on self-report, and the saliva test does not follow. On quality of life, on mood, on sleep, nothing moves, and one patient in four stops for an adverse event. I take it that the lead is serious and that I still have nothing to promise a patient. ”
What this means in practice: faced with any positive trial, look for whether the end point is self-reported or objective, and whether an independent objective measure exists inside the trial. Here it does exist, and it does not confirm.
What you can do with this on Monday morning. Practice does not change this week. The conversation, on the other hand, is already arriving, carried by press coverage.
- Set out the regulatory position without approximation. Mirtazapine is licensed as an antidepressant, for the treatment of major depressive episodes in adults. Methamphetamine use disorder is not among its approved indications, and, as the authors state, no medicine at all is approved for that disorder. The exact wording of the approved indication and the approved dose range are set nationally: they are found in the summary of product characteristics or the product label in force where you practise, not in a trial report.
- Say clearly what that situation implies. Prescribing mirtazapine in methamphetamine use disorder would today be off-label use. PEB neither recommends it nor advises against it: that is not the role of a journal, and it is not what these data allow. A prescriber who considered it takes on the responsibility, and it falls to them to check the obligations that attach to off-label prescribing where they practise, obligations this analysis has not verified at source. For information, the investigators announce in their published protocol that a favourable result will serve as the basis for an application for methamphetamine use disorder to be included as a therapeutic indication. At the date this trial was published, no medicine holds an approval in this indication.
- Do not confuse this study with a depressive indication. The point is counter-intuitive and deserves to be said out loud: in this trial, mirtazapine did not improve depressive symptoms. Justifying a prescription by the mood comorbidity of a patient who uses methamphetamine falls under the approved indication of the drug and is judged on its own criteria, not on this work.
- Answer the patient who has read the press with figures, not with a position. Two fewer days of use per month. No demonstrated effect on sleep, mood or quality of life. Close to one patient in four stops for an adverse event, against roughly one in seven on placebo. Drowsiness in close to one patient in two. This is a serious lead, it is not an available treatment.
- Document and monitor, if the question arises all the same. Weight and metabolic data at initiation and then regularly, drowsiness and its effect on driving or work, and a trial duration defined in advance rather than treatment left to run. Continuing psychosocial interventions falls under treatment guidelines, outside this trial: the published abstract does not state what support was delivered in each arm, and nothing in these data suggests scaling it back. On what those interventions do and do not achieve, see the Cochrane review of psychosocial interventions in stimulant use disorder, retention against abstinence.
- Take away the reading grid. Self-reported end point against objective end point, recognisable adverse effects and integrity of the blind, the bound of the confidence interval nearest the null rather than the value of p. These three reflexes serve well beyond addiction medicine.
The epidemiological question
The idea that methamphetamine use is rising circulates widely and often serves to justify the local applicability of this kind of result. It has to be handled with the same caution as the trial itself, because most national indicators do not separate methamphetamine from the other amphetamines, neither for prevalence of use in the general population nor for seizures. Where a source does not isolate the substance, no figure specific to methamphetamine can be drawn from it.
At European level, the 2026 report of the European Union Drugs Agency, covering 2024 data, places treatment demand related to methamphetamine in four countries: Czechia, Germany, Slovakia and Türkiye alone account for 92% of the roughly 14,500 patients entering treatment for methamphetamine in the countries the report covers. Seizures declared by member states, on the other hand, rose sharply, from 1.8 tonnes in 2023 to 6.1 tonnes in 2024. Consulted on 10 August 2026.
The conclusion to draw is a measured one. A rise in availability across Europe is documented by the seizures member states declare. A rise in use is a separate claim, and in most countries the indicators that would support it are not available in a form that isolates methamphetamine. Absence of a figure is neither proof that there is no problem, nor proof that there is one. For a clinician working outside the few countries where treatment demand is concentrated, the immediate reach of this trial therefore remains limited to individual situations, and its main significance is scientific: the first phase 3 demonstration of mirtazapine in this disorder.
Frequently asked questions
Is this the first effective drug treatment for methamphetamine use disorder?
No. It is the first phase 3 trial of mirtazapine in this indication, and its primary end point is positive. It is neither the first positive trial, nor the first phase 3 trial conducted in this disorder. Two phase 2 randomised trials, published in 2011 and 2020, had already reported a reduction in use on mirtazapine, in San Francisco, among men who have sex with men for the first, among cisgender men and transgender women who have sex with men for the second, with high HIV risk in both cases. And the ADAPT-2 trial, phase 3, had reported a positive result in 2021 on a biological end point with naltrexone combined with bupropion, in 403 participants. What remains accurate, and what the abstract of the article restates, is that no medicine holds an approval in this indication at the date of publication.
Can mirtazapine be prescribed to a patient who uses methamphetamine?
Not in this indication under a marketing authorisation. Mirtazapine is licensed for the treatment of major depressive episodes in adults, and no medicine is approved for methamphetamine use disorder. A prescription in this disorder would be off-label use, with the obligations and the responsibility that attach to it where you practise.
Why insist on the primary end point being self-reported?
Because mirtazapine produces recognisable adverse effects, drowsiness in 47% of patients against 33% on placebo, weight gain in 10% against 3%. A patient can therefore guess their group. When the end point rests on what that same patient reports, the risk of a bias oriented in favour of treatment is not theoretical. And in this trial, the only objective measure available, oral fluid, does not confirm the self-reported result.
Two fewer days of use per month, is that clinically important?
Nobody can answer that with data. No validated threshold of clinical relevance exists for this end point. Two things can only be observed. The reduction is not accompanied, in this trial, by any significant improvement in quality of life, mood or sleep. And the trial had been sized to detect a reduction of roughly five days out of 28, with 90% power: the effect obtained is clearly below its own design target. Over twelve weeks, the improvement stays quantitative.
How widespread is methamphetamine use disorder?
Less evenly than the coverage suggests, and it cannot be quantified everywhere. Many national sources do not separate methamphetamine from the other amphetamines, neither for prevalence of use nor for seizures, which are published for the substances combined. At European level, treatment demand remains concentrated in four countries, while seizures rose sharply in 2024. Availability is rising; use in a given country is often not measured separately.
Would a longer treatment, or a higher dose, do better?
That is a hypothesis, not a result. The trial tested 30 mg daily for twelve weeks, and nothing in these data informs any other dose or any other duration. It is worth recalling that the approved dose range for mirtazapine applies to the depressive indication, which is the only approved one.
Annotated bibliography
McKetin R, Shoptaw S, Saunders L, Nguyen L, Clare PJ, Dore GJ, Turner A, Dean OM, Kelly PJ, Arunogiri S, et al., Berk M (2026). Mirtazapine for Methamphetamine Use Disorder: A Randomized Clinical Trial. JAMA Psychiatry, 83(6), 581-589. DOI 10.1001/jamapsychiatry.2026.0159 · PMID 41920558. Source study analysed here. Phase 3 trial, six Australian sites, 344 randomised, 339 treated. Registration ACTRN12622000235707. Pagination and volume verified on 10 August 2026 on the Europe PMC record. The interest declarations and the funding statement published with the article were not consulted. Only the published abstract was consulted, the full text was not. An erratum has been published on the visual abstract, JAMA Psychiatry 2026;83(6):658, DOI 10.1001/jamapsychiatry.2026.1470 · PMID 42234442. It concerns that illustration alone, and no figure in this analysis is drawn from it.
McKetin R, Degan TJ, Saunders L, et al., Berk M (2024). A phase 3 randomised double-blind placebo-controlled trial of mirtazapine as a pharmacotherapy for methamphetamine use disorder: a study protocol for the Tina Trial. Trials, 25(1), 408. DOI 10.1186/s13063-024-08238-y · PMID 38907288. Published protocol of the trial, consulted in full text. Source of the methodological items absent from the abstract of the article: the self-reported nature of the primary end point, eligibility from 18 to 65 years, 1:1 allocation stratified by site, sex and PHQ-9 score, take-home medication, a dosing schedule of 30 mg in the evening without titration with a possible reduction to 15 mg and a taper to 15 mg for 28 days from week 12, adherence monitoring with MEMS caps, a target of 340 participants sized to detect a minimum rate ratio of 0.75 with 90% power, a mixed model with a group by time interaction, a primary intention-to-treat analysis on non-imputed data, sensitivity analyses with imputation and a per-protocol analysis. Declared funding: the Australian Medical Research Future Fund, grant no. 2007155, with no role of the funder in the conduct of the study. It is also the source of the priority claim, which is restricted to mirtazapine: “the first phase 3 placebo-controlled randomised trial to examine the effectiveness and safety of mirtazapine as an outpatient pharmacotherapy for methamphetamine use disorder”.
Coffin PO, Santos GM, Hern J, Vittinghoff E, Walker JE, Matheson T, Santos D, Colfax G, Batki SL (2020). Effects of Mirtazapine for Methamphetamine Use Disorder Among Cisgender Men and Transgender Women Who Have Sex With Men: A Placebo-Controlled Randomized Clinical Trial. JAMA Psychiatry, 77(3), 246-255. DOI 10.1001/jamapsychiatry.2019.3655 · PMID 31825466. Enlarged replication trial, 120 participants included, 115 cisgender men and 5 transgender women. Dosing of 15 mg in the first week then 30 mg daily through to week 24, a titration that the Tina trial did not adopt. Relative risk of positive urine tests of 0.67 at week 12 (95% CI 0.51 to 0.87), with an objective end point this time. Measured adherence was very low and comparable in the two arms between weeks 2 and 12, 38.5% on mirtazapine against 39.5% on placebo, but divergent between weeks 13 and 24, 28.1% against 38.5%: the equivalence of adherence therefore holds only for the first period. A very specific population, not transposable to a general caseload.
Colfax GN, Santos GM, Das M, Santos DM, Matheson T, Gasper J, Shoptaw S, Vittinghoff E (2011). Mirtazapine to reduce methamphetamine use: a randomized controlled trial. Archives of General Psychiatry, 68(11), 1168-1175. DOI 10.1001/archgenpsychiatry.2011.124 · PMID 22065532. The first randomised trial of the drug in this indication, 60 participants, all men who have sex with men, recruited in San Francisco between 2007 and 2010. Mirtazapine 30 mg daily from the outset, without titration. Relative risk of 0.57 (95% CI 0.35 to 0.93; P = .02) on urine tests. A small sample, but an objective end point and adherence measured at 48.5% by an electronic device.
Trivedi MH, Walker R, Ling W, Dela Cruz A, Sharma G, Carmody T, Ghitza UE, Wahle A, Kim M, Shores-Wilson K, et al., Shoptaw S (2021). Bupropion and Naltrexone in Methamphetamine Use Disorder. New England Journal of Medicine, 384(2), 140-153. DOI 10.1056/NEJMoa2020214 · PMID 33497547. The ADAPT-2 trial, NIDA CTN-0068 protocol, registered as NCT03078075 and declared phase 3 on that registry. 403 participants enrolled in the first stage, sequential parallel comparison design, extended-release injectable naltrexone combined with oral bupropion. Biological primary end point: at least three of four negative urine samples at the end of the stage. Weighted response of 13.6% against 2.5% on placebo, a difference of 11.1 points (P < 0.001). It is this trial that rules out writing that Tina is the first phase 3 trial conducted in this disorder.
Naji L, Dennis B, Rosic T, Wiercioch W, Paul J, Worster A, Thabane L, Samaan Z (2022). Mirtazapine for the treatment of amphetamine and methamphetamine use disorder: A systematic review and meta-analysis. Drug and Alcohol Dependence, 232, 109295. DOI 10.1016/j.drugalcdep.2022.109295 · PMID 35066460. The state of the evidence before Tina: two trials, 180 participants in total, a probably small reduction in use at twelve weeks with a relative risk of 0.81 whose 95% confidence interval (0.63 to 1.03) crosses the null, in 133 participants, at moderate certainty. This pooled value does not rest on the same event definition as the 0.57 and 0.67 reported by each trial, which follow their own conventions for analysing urine tests: the three figures do not compare term for term. Neither retention in treatment nor depressive symptoms were improved, which the phase 3 trial reproduces for depression.
European Union Drugs Agency. Synthetic stimulants, the current situation in Europe, European Drug Report 2026. euda.europa.eu. Consulted on 10 August 2026, data for the year 2024. Source of the concentration of treatment demand in four countries, 92% of the roughly 14,500 patients entering treatment, and of the change in seizures from 1.8 tonnes in 2023 to 6.1 tonnes in 2024. An agency report based on national indicators of variable quality from one country to another.
What was consulted. References verified on Europe PMC on 10 August 2026, epidemiological data verified on 10 August 2026 on the site of the European Union Drugs Agency, phase and sample size of the ADAPT-2 trial verified on ClinicalTrials.gov. The published protocol of the Tina trial was consulted in full text, as were the publications of Colfax 2011, Coffin 2020 and Naji 2022. Of the source study, only the published abstract could be consulted: the full text, although open access, was not reachable by the tools used on 10 August 2026. This analysis underwent an independent double reading.