Published on 16 September 2026
Brexpiprazole plus sertraline in PTSD: fixed-dose phase 3 trial does no better than sertraline
In brief
A phase 3 trial randomised 553 adults with post-traumatic stress disorder, across 95 selected US sites of which 91 actually enrolled, to one of two fixed doses of brexpiprazole added to sertraline 150 mg per day, or to sertraline alone. On the primary outcome, change in CAPS-5 total score from week 1 to week 10, the pre-specified comparison of the average effect of the two doses is null (p=0.90), and the testing hierarchy stops there. Taken separately, the two doses do not separate either: a difference of +1.03 points for 2 mg (95% confidence interval −2.09 to 4.16; p=0.52) and of −0.71 points for 3 mg (−3.88 to 2.46; p=0.66), a negative sign favouring the combination. Effect sizes range from −0.11 to 0.17 across all outcomes, with one exception, psychosocial functioning in the 3 mg arm, where d reaches 0.36. The trial is soundly conducted and its negative result is published without spin, which deserves notice when eight of the nine authors are employees of the two sponsors. Its limitations are not primarily statistical: the comparator is sertraline alone rather than placebo, both drugs are started together in patients who were not already on effective sertraline, responders to the placebo run-in are excluded from the efficacy analysis, and the doses were fixed whereas the two earlier flexible-titration trials were positive. The confidence intervals exclude the targeted 5-point gain; they do not establish equivalence, and they do not exclude a small residual benefit.
The context
Post-traumatic stress disorder is one of the few psychiatric disorders in which the validated pharmacopoeia fits in a few lines, and in which partial response is the rule rather than the exception. In the United States, sertraline and paroxetine have remained the only two drugs approved for this indication for more than twenty years. Adding a second-generation antipsychotic is one of the most frequently improvised responses, on the strength of small trials and of a habit carried over from treatment-resistant depression.
One point of reading needs to be made at the outset, because everything else depends on it. The trial analysed here is not an augmentation trial in patients already stabilised on a reuptake inhibitor: patients receiving sertraline at an adequate dose and for an adequate duration, more than 50 mg per day for at least eight weeks, were explicitly excluded, ongoing psychotropic drugs were stopped during screening, and the two drugs were started together after one week of placebo. The question asked is therefore that of a combination from the outset, not of what to add for a partial responder.
The brexpiprazole development programme gave this hypothesis some initial support: two earlier flexible-titration trials, numbered 061, a full-factorial phase 2 trial that included a brexpiprazole monotherapy arm, and 071, a phase 3 trial, had concluded that the combination was superior to sertraline alone. The trial analysed here, numbered 072 and run in parallel with 071, is the expected confirmation with a fixed-dose design. It does not confirm.
The mechanism
What was measured, and what was not
The pharmacological hypothesis behind the combination belongs to general knowledge of the drug, not to the results of this trial. It is recalled here to situate what the trial was able to test.
| Item | What is known |
|---|---|
| Pharmacological hypothesis | Action on the three monoamines added to reuptake inhibition |
| FindingA pharmacological reminder, outside the scope of the trial. The authors invoke dysregulation of the noradrenergic, serotonergic and dopaminergic systems, and an action of brexpiprazole on these three systems, where the reuptake inhibitor acts mainly on one. The hypothesis is plausible; it has never been validated by an intermediate marker in post-traumatic stress disorder. | |
| What the trial measures | Clinical rating scales, at ten and twelve weeks |
| FindingThe clinician-rated CAPS-5 as primary outcome, then the Clinical Global Impression of severity and psychosocial functioning as key secondary outcomes, then so-called “other” outcomes: a self-report symptom questionnaire, anxiety and depression, CAPS-5 cluster scores, response rates. No biological surrogate outcome is involved, which is an advantage here: the primary outcome is the symptom itself. | |
| What is not reported | Receptor occupancy and plasma concentrations |
| FindingThe publication reports neither plasma levels nor any pharmacokinetic analysis. Nothing in it therefore shows whether the fixed doses of 2 and 3 mg produced an exposure comparable to that reached by titration in the earlier trials. As the supplementary digital content could not be consulted, it cannot be asserted that no such data exist: we can only observe that they do not appear in the article. | |
The study at a glance
| Population, intervention, comparator, outcomes | |
|---|---|
| Population | |
| PAdult outpatients aged 18 to 65 years, DSM-5 post-traumatic stress disorder confirmed by the MINI, symptoms for at least six months, CAPS-5 score of at least 33 at screening and at baseline. Index trauma occurring after the age of 16 and within the past nine years. A current major depressive episode, treatment with sertraline that was already effective, and compensation or litigation related to the disorder were all grounds for exclusion. Mean age 37.0 years, 74.0% women, 67.8% white participants. | |
| Intervention | |
| IBrexpiprazole 2 mg per day combined with sertraline 150 mg per day, or brexpiprazole 3 mg per day combined with sertraline 150 mg per day. Fixed target doses, reached through a mandatory titration over two to three weeks, with no individual adjustment permitted thereafter. | |
| Comparator | |
| CSertraline 150 mg per day combined with brexpiprazole placebo. The comparator is therefore an active treatment, not a pure placebo: all three arms receive sertraline. | |
| Primary outcome | |
| OChange in CAPS-5 total score from randomisation, at week 1, to week 10, analysed with a mixed model for repeated measures. | |
| Secondary outcomes | |
| OOnly two key secondary outcomes are included in the testing hierarchy, in this order: Clinical Global Impression of severity, from week 1 to week 10, then the Brief Inventory of Psychosocial Functioning, from baseline to week 12. The self-report post-traumatic symptom questionnaire and the Hospital Anxiety and Depression Scale are “other” outcomes, tested at the nominal 0.05 level with no correction for multiplicity. | |
| Design and numbers | |
| DPhase 3, multicentre, randomised 1:1:1, double-blind trial with three parallel arms, level of evidence 1b. Twelve weeks in all: one week of placebo, then eleven weeks of randomised treatment. 1821 people assessed at screening, 591 entered the placebo run-in, 553 randomised (191, 185 and 177), 537 in the safety sample and 388 in the efficacy sample. Recruitment from October 2019 to August 2023. |
Quality control
| Item | Judgement |
|---|---|
| Design and conduct | Double-blind phase 3, 553 randomised |
| FindingComputer-generated randomisation in fixed blocks of 3, stratified by site and by enrichment status. Participants, investigators and sponsor staff, analysts included, were blinded. Participants and site teams were not even told of the existence of the placebo run-in, the timing and nature of randomisation, or the timing of the primary outcome. Analysis by mixed model for repeated measures, reporting in line with the CONSORT guidelines. | |
| Reporting of the result | Negative result published without spin |
| FindingThe testing hierarchy is respected, and the two nominally significant differences, on psychosocial functioning and on response rate in the 3 mg arm, are presented with their limitations rather than showcased. In a field where publication bias against negative trials remains massive, this needs to be said. The authors’ conclusion nonetheless pushes the overall reading of the programme in a direction favourable to the product. | |
| Independence | Eight of nine authors employed by the sponsors |
| FindingSix authors are or were employees of Otsuka, two of Lundbeck. Only the first author is external, with declared ties that include advisory board fees from Otsuka. The sponsors state that they were involved in the design, conduct, collection, analysis and interpretation of the data, as well as in the preparation and approval of the article and the decision to submit it. Writing support was funded by them. The usual direction of allegiance bias would favour the product, which makes the negative result more credible, but does not remove the need to check the protocol and the statistical analysis plan. | |
| Dosing regimen | Fixed doses, earlier trials flexible-dose |
| FindingThe two earlier positive trials used flexible titration. The authors report having run multiple post hoc analyses to explain the divergence, without finding its cause. The question remains open: individual adjustment of exposure, different patient selection, regression to the mean, or simply chance across the three trials. | |
| Population selection | Enrichment applied to the analysis |
| FindingA point not to be confused: responders to the placebo week were not kept out of randomisation, they were excluded from the efficacy sample. That sample retains only participants who still had a CAPS-5 score of at least 27 at week 1 and less than 50% improvement since baseline, that is 388 of the 553 randomised. The procedure is legitimate and increases the trial’s sensitivity, and randomisation was indeed stratified on this status. It does, however, produce a population that is harder to treat than that of an ordinary outpatient clinic. | |
| External validity | United States only, aged 18 to 65 |
| FindingRecruitment exclusively in the United States. Contrary to what the North American literature on this disorder might suggest, the population is not military: the index trauma was war or combat-zone exposure in 6 of 553 participants, about 1%, whereas the protocol allowed up to 20%. Assault predominates, from 32.4% to 39.5% depending on the arm, followed by sexual trauma, from 19.2% to 22.7%, exposure to sudden death, from 15.8% to 21.1%, and road traffic accidents, around 10%. Excluding patients receiving compensation or involved in litigation, and those with a current major depressive episode, leaves out a large share of the patients seen in routine practice. | |
The findings
| Outcome | Result |
|---|---|
| First-rank comparison | Average effect of the two doses, p=0.90 |
| FindingThe pre-specified procedure first tested the average effect of the two brexpiprazole arms against sertraline alone, before moving down to the doses taken separately. This test is null, and it is the one that halts the hierarchy. | |
| Primary outcome, brexpiprazole 2 mg | +1.03 CAPS-5 points (−2.09 to 4.16), p=0.52 |
| FindingThe sign is positive, that is, against the combination: this arm improves very slightly less than sertraline alone. The most favourable bound of the confidence interval, −2.09 points, remains far from the 5 points targeted. | |
| Primary outcome, brexpiprazole 3 mg | −0.71 CAPS-5 points (−3.88 to 2.46), p=0.66 |
| FindingThe sign reverses, this time in favour of the combination, with an equally negligible magnitude and a favourable bound at −3.88 points. Two estimates of opposite sign for two doses expected to produce an increasing effect: this is the pattern expected from an absence of effect, not from an effect too small to be detected. | |
| Size of the improvement | −16.5, −18.3 and −17.6 CAPS-5 points |
| FindingAll three arms improve clearly and to an overlapping degree, from a mean score of 37.9 to 39.3 at randomisation. On top of this comes a fall of 8.8 points during the placebo run-in week alone. This improvement cannot be attributed to sertraline: without a pure placebo arm, spontaneous course, regression to the mean and the non-specific effects of follow-up cannot be separated from the effect of the drug. | |
| Effect sizes | Cohen’s d from −0.11 to 0.17, bar one outcome |
| FindingOn the primary outcome, its four cluster scores, the Clinical Global Impression, the self-report symptom questionnaire and the two anxiety and depression subscales, no effect size reaches 0.20, the conventional threshold for a small effect. The only value that stands out is psychosocial functioning at 3 mg, d=0.36. Consistency across outcomes is more informative here than any single value taken in isolation. | |
| Testing hierarchy | Stopped at the first level |
| FindingAs the primary outcome was not met, the pre-specified procedure halts: the two key secondary outcomes that followed, Clinical Global Impression and then psychosocial functioning, lose their confirmatory status. The p values that appear afterwards are nominal and unprotected against multiplicity. | |
| Psychosocial functioning, 3 mg arm | −8.83 points (−15.82 to −1.85), nominal p 0.013 |
| FindingThis is the last rung of the hierarchy, stripped of its protection when the hierarchy stopped, not an outcome that sat outside the hierarchy from the start. The scale runs from 0, minimal impairment, to 100: the negative value therefore favours the combination. It rests on a self-rating, measured from baseline to week 12 rather than week 10, in smaller numbers than the primary outcome, 94 to 98 per arm. It is not replicated in the 2 mg arm, where the difference is −4.16 points, p=0.23. It justifies a hypothesis, not a decision. | |
| CAPS-5 response rate | 69.1% versus 60.0%, nominal p 0.043 at 3 mg |
| FindingThe trial’s second nominally significant result, on an “other” outcome, defined as at least 30% improvement from week 1 and analysed with last observation carried forward. The response ratio is 1.21, confidence interval 1.00 to 1.46, with the lower bound right at the threshold. At 2 mg, 62.9% and p=0.52. No correction for multiplicity was applied to these outcomes. | |
| Tolerability | No new safety signal reported |
| FindingTreatment-emergent adverse events in 51.4%, 48.3% and 51.2% of participants on 2 mg, 3 mg and sertraline alone. Discontinuations due to adverse events: 2.7%, 4.4% and 4.7%. One death, by accidental drowning in a bathtub, judged unrelated to treatment. The commonest effects, nausea, headache and diarrhoea, are those of sertraline and are evenly distributed across the arms. | |
| Weight and motor effects | Weight gain and akathisia more frequent with the combination |
| FindingMean weight change of +0.8 kg at 2 mg, +0.9 kg at 3 mg and −0.4 kg on sertraline alone. A gain of at least 7% of body weight occurred in 9.2%, 8.0% and 4.1% of participants. Extrapyramidal adverse events: 9.2%, 6.7% and 4.7%. Akathisia: 3.8%, 3.3% and 1.2%. No new signal therefore does not mean no adverse effects, and eleven weeks say nothing about long-term metabolic or motor risk. | |
Critical appraisal
| Domain | Judgement |
|---|---|
| Randomisation process | Fixed blocks of 3, stratified on two factors |
| FindingComputer-generated sequence, supplied by the sponsor, blinded allocation, stratification by site and by enrichment status. The table of baseline characteristics shows no notable imbalance between the three arms: age, sex, ethnicity, time since trauma, nature of the index trauma, treatment history and baseline scores overlap. | |
| Deviations from intended protocol | Blinding maintained, doses not modifiable |
| FindingThe fixed-dose design removes the risk of unblinding through titration, but rules out individual adjustment. This choice protects internal validity at the direct expense of external validity. Two amendments are declared, one relabelling the non-key secondary outcomes, the other reducing the enrolment target from 733 to 585 participants because of the pandemic. | |
| Missing data | About 34% study withdrawals |
| FindingCompletion rates are 66.0%, 66.5% and 65.5%, and therefore balanced. This level is usual for trials of this length in this disorder, but it remains high in absolute terms. The mixed model assumes data missing at random conditional on the observations, an assumption that cannot be checked. The loss is even greater for psychosocial functioning, measured at week 12. | |
| Measurement of the outcome | Clinician-rated CAPS-5, blinded |
| FindingThe CAPS-5 is the reference instrument in the field, administered by a clinician kept blind. The version used after baseline covers the past week. The starting point for change is indeed randomisation at week 1, not baseline: the 8.8 points gained during the placebo week are therefore not counted in the reported magnitude. | |
| Selection of the reported result | Hierarchy pre-specified and respected |
| FindingThe protocol and statistical analysis plan are deposited and publicly available, and the trial is registered as NCT04174170. The two nominally positive results are presented as such, which is the expected conduct and one rarely seen. | |
| Competing interests | Sponsors own the data |
| FindingEight of the nine authors are employees of the two sponsors, Otsuka and Lundbeck, which fund the trial and state that they were involved at every stage up to the decision to submit. Formally, the risk of allegiance is at its maximum. Here it turns into an argument for credibility, since the result runs against the commercial interest of the product. | |
Level of evidence
What the trial demonstrates: in a population of US adults whose efficacy analysis excludes responders to one week of placebo, starting brexpiprazole at fixed doses of 2 or 3 mg per day together with sertraline 150 mg per day does not produce, at ten weeks, the 5-point CAPS-5 gain for which the trial was sized. The sample size calculation aimed for at least 80% power for this 5-point difference, standard deviation 14, with 375 evaluable participants, and at least 90% power for the first-rank comparison of the average effect of the two doses. The two estimates are of opposite sign, of negligible magnitude, and their confidence intervals exclude the targeted gain, the most favourable bound being 3.88 points at 3 mg and 2.09 points at 2 mg. Confidence on this point is high.
What the trial does not demonstrate, and this is where symmetry must be held: the absence of a significant difference is not evidence of equivalence. The trial was never built for that; it has neither a non-inferiority margin nor a pre-specified equivalence test, and its confidence intervals leave room for a benefit of a few points, in either direction. What can be said at this stage is that the targeted additional benefit is ruled out and that a small residual benefit is not. What the trial suggests, and no more, are two concordant signals in the 3 mg arm, on psychosocial functioning and on response rate, both nominal and unprotected against multiplicity, the second of which barely crosses the threshold.
What the trial cannot rule out: an effect in patients who improve from the outset, excluded from the analysis by the enrichment criterion, and therefore in a large share of everyday outpatients; an effect obtained through individual titration, since the two earlier flexible-dose trials were positive and the authors’ post hoc analyses did not find the source of the divergence; an effect beyond ten weeks; an effect in depressed patients, excluded by the protocol; an effect in subgroups that the trial’s power does not allow to be explored. Finally, the three-arm design says nothing about the efficacy of brexpiprazole used alone, which was not tested here, nor about that of sertraline, present in all three arms. This is not a placebo-controlled trial, and the improvement common to the three arms cannot be credited to any treatment.
Finally, the following belong to expert opinion, not to this trial: the hypothesis that flexible titration explains the divergence between the three trials of the programme, the authors’ reading that the programme as a whole points to a consistent effect of the combination, and the idea that this result would hold for all second-generation antipsychotics used as add-ons in this indication. None of the three is established by the data presented.
The colleague test
What an experienced colleague would say if you put this study to them in two minutes, between two consultations.
“Right, I am not starting this combination in PTSD, that much is settled. A clean phase 3, more than five hundred patients, and both curves sit right on top of sertraline alone. What bothers me is that the two earlier trials, with flexible dosing, were positive, and nobody can say why. So I do not prescribe it, but I do not close the file either.”
What this means in practice: there is no reason today to start fixed-dose brexpiprazole together with a reuptake inhibitor from the outset in this indication. That is not a demonstration that the drug is inert in the disorder: it is a demonstration that this regimen, in this population, adds nothing measurable at ten weeks.
What you can do with this
- Do not start the fixed-dose combination. This trial provides no argument for combining brexpiprazole 2 or 3 mg per day with sertraline from the outset. The reasoning carried over from treatment-resistant depression does not hold here.
- Do not carry this result over to the partial responder. Patients already on effective sertraline were excluded from the trial, which therefore did not test augmentation in a patient stabilised on a reuptake inhibitor. There is no argument for that practice either; there is an absence of data.
- Go back to the fundamentals before adding a drug. The dose and duration actually received, adherence, comorbid addiction or depression, sleep, and above all access to structured trauma-focused psychotherapy, which remains the backbone of international guidelines.
- Know what to say to a patient already on this combination. A stable patient should not be destabilised by the reading of a negative trial. The reasonable course is to reassess the perceived individual benefit, to discuss tapering if nothing can be identified, and not to stop in a hurry.
- Monitor what remains true whatever happens. In a patient exposed to a second-generation antipsychotic, metabolic, weight and motor monitoring does not depend on the efficacy observed. In this very trial, weight gain of at least 7% and extrapyramidal effects were twice as frequent with the combination.
- Keep the methodological lesson. Two positive trials followed by a negative confirmatory trial is the normal course of a development programme, not an anomaly. It calls for caution in the face of any early signal, including when it comes from randomised trials.
- The course of action is set out in the NICE decision tree for depression in adults.
Regulatory context, as of 13 August 2026 and to be checked before any reuse. Marketing authorisation, actual marketing and reimbursement are three separate questions, they are settled country by country, and an answer to one is not an answer to the other two. In the European Union, brexpiprazole has held a centralised marketing authorisation since 26 July 2018, and its current indication there is the treatment of schizophrenia in adults and in adolescents aged 13 years and older; it has no indication in post-traumatic stress disorder in the European Union. For the United States, the publication itself states: “At the time of writing, brexpiprazole in combination with sertraline is not approved by the FDA for the treatment of PTSD.” Wherever prescribing outside an authorised indication is legally possible, this trial provides no grounds for considering it in this indication.
Frequently asked questions
Does this trial show that brexpiprazole does not work in PTSD?
No. It establishes that combining it from the outset, at fixed doses of 2 or 3 mg per day, with sertraline 150 mg per day brings no measurable benefit at ten weeks, in a population whose analysis excludes responders to one week of placebo. The drug used alone was not tested here, nor was dose-adjusted administration, and an absence of significant difference is never to be read as evidence of absence of effect.
Why was the brexpiprazole plus sertraline trial not placebo-controlled?
Because the question asked concerns the combination, not the efficacy of brexpiprazole in its own right. All three arms receive 150 mg of sertraline. The consequence has to be followed through: the improvement of 16.5 to 18.3 points seen in every arm does not demonstrate the efficacy of sertraline, since spontaneous course and the non-specific effects of follow-up are not isolated. The preceding placebo week had in fact already lowered the score by 8.8 points.
Why were the earlier brexpiprazole trials in PTSD positive when this one was not?
No explanation is established. The authors state that they ran multiple post hoc analyses, on methodology as well as on patient characteristics, without finding a clear cause. The hypotheses put forward, individual titration allowing an effective exposure to be reached, a difference in population, regression to the mean or simple variability between trials, are plausible and unresolved. It is an open question, and presenting it as settled in either direction would be a misreading.
Is the psychosocial functioning benefit with brexpiprazole plus sertraline clinically usable?
Not as an argument for prescribing. It sat on the last rung of the testing hierarchy, which stopped at the primary outcome: its p value is therefore no longer protected against multiplicity. It rests on a self-rating measured at week 12, in reduced numbers, and it is not replicated in the 2 mg arm. It can legitimately feed a hypothesis for a later dedicated trial, nothing more.
Should brexpiprazole be stopped in a patient doing well on the combination?
A group trial does not dictate individual management. The reasonable approach is to reassess with the patient what can be attributed to the treatment, to consider gradual tapering if no benefit can be identified, taking into account the risk of adverse effects from a treatment whose additional benefit is not demonstrated, and to maintain metabolic monitoring for as long as exposure continues.
Annotated bibliography
Source study. Davis LL, Behl S, Lee D, Zeng H, Skubiak T, Weaver S, Hefting N, Groes Larsen K, Hobart M. Fixed-Dose Brexpiprazole and Sertraline Combination Therapy for the Treatment of Posttraumatic Stress Disorder: A Phase 3, Randomized Trial. Journal of Clinical Psychopharmacology, 2025; 45(6): 580-589. DOI 10.1097/JCP.0000000000002076. PMID 40982703. Trial 072, registered as NCT04174170, randomised double-blind phase 3 trial, three parallel arms, 553 patients randomised across 95 selected US sites. Contribution: a negative result with a high level of evidence on a clinical development question, published without minimisation despite industry sponsorship. Limitations: recruitment exclusively in the United States, exclusion of early responders from the efficacy analysis, fixed doses not matching the earlier trials, data controlled by the sponsors.
Trial 071, flexible-titration phase 3. Davis LL, Behl S, Lee D, et al. Brexpiprazole and Sertraline Combination Treatment in Posttraumatic Stress Disorder: A Phase 3 Randomized Clinical Trial. JAMA Psychiatry, 2025; 82(3): 218-227. DOI 10.1001/jamapsychiatry.2024.3996. Difference of −5.59 CAPS-5 points in favour of the combination, 95% confidence interval −8.79 to −2.38, p<0.001. Contribution: the twin of the present trial, run in parallel, with the same primary outcome and an otherwise identical design, dosing aside. Limitations: the difference in results from trial 072 remains without an established explanation, and the only documented design difference is the dosing regimen.
Trial 061, full-factorial phase 2. Hobart M, Chang D, Hefting N, Davis LL. Brexpiprazole in Combination With Sertraline and as Monotherapy in Posttraumatic Stress Disorder: A Full-Factorial Randomized Clinical Trial. The Journal of Clinical Psychiatry, 2025; 86: 24m15577. DOI 10.4088/JCP.24m15577. Difference of −5.08 CAPS-5 points, 95% confidence interval −8.96 to −1.20, p=0.011. Contribution: the only trial in the programme with a brexpiprazole monotherapy arm, and the only one whose primary analysis is not enriched. Limitations: phase 2, smaller sample, a wider window for time since trauma, fifteen years against nine, which limits direct comparability with the two phase 3 trials.
Primary measurement instrument. Weathers FW, Bovin MJ, Lee DJ, et al. The Clinician-Administered PTSD Scale for DSM-5 (CAPS-5): Development and initial psychometric evaluation in military veterans. Psychological Assessment, 2018; 30(3): 383-395. PMID 28493729. The trial uses the “past-month” and “past-week” versions of the instrument, distributed by the National Center for PTSD. Contribution: the reference instrument in the field, with good psychometric properties, and clinician rating that is less sensitive to expectation than self-report. Limitations: initial validation was carried out in veterans, whereas the trial population is 99% civilian; and there is no consensus threshold of clinical relevance for change in total score, which makes the interpretation of a difference of a few points dependent on convention rather than data.
Regulatory documents. European public assessment report and summary of product characteristics for brexpiprazole, European Medicines Agency, marketing authorisation of 26 July 2018. Contribution: they set the scope of the authorised indications in the European Union, the only enforceable frame there in practice. Limitations: they do not settle actual marketing, which depends on separate decisions, and they concern schizophrenia, not the indication discussed here.
What was consulted. References verified on 13 August 2026 against the published version, and against the PubMed and publisher records of the references cited. The supplementary digital content of the source study, tables S1 to S3 and figures S1 and S2, could not be consulted: the links in the publication lead to it through restricted access. None of the data in this analysis come from it. Article subject to an independent double reading.